monte carlo methods for partial differential equations: a ... · monte carlo methods for partial...
TRANSCRIPT
![Page 1: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/1.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Monte Carlo Methods for Partial DifferentialEquations: A Personal Journey
Prof. Michael Mascagni
Department of Computer ScienceDepartment of Mathematics
Department of Scientific ComputingGraduate Program in Molecular Biophysics
Florida State University, Tallahassee, FL 32306 USAE-mail: [email protected] or [email protected]
URL: http://www.cs.fsu.edu/∼mascagniIn collaboration with Drs. Marcia O. Fenley, Nikolai Simonov and Alexander Silalahi, Robert
Harris, and Messrs. Travis Mackoy, and James McClain
Research supported by ARO, DOE, NATO, and NSF
![Page 2: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/2.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
IntroductionA Little History on Monte Carlo Methods for PDEsSome Examples Using This for Computing Elliptic Problems
The Walk on Spheres MethodProblems in Electrostatics/MaterialsVarious Acceleration Techniques for Elliptic PDEs
Mathematical ModelElectrostatic Potential and EnergyThe Feynman-Kac Formula
Fast Exit Point Calculations‘Walk-on-Spheres’ AlgorithmWalk-in-SubdomainsMonte Carlo Treatment of Boundary Conditions
Monte Carlo EstimatesMonte Carlo EstimatesComputational GeometryCorrelated and Uncorrelated Sampling
Computational ResultsConclusions and Future Work
![Page 3: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/3.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
IntroductionA Little History on Monte Carlo Methods for PDEsSome Examples Using This for Computing Elliptic Problems
The Walk on Spheres MethodProblems in Electrostatics/MaterialsVarious Acceleration Techniques for Elliptic PDEs
Mathematical ModelElectrostatic Potential and EnergyThe Feynman-Kac Formula
Fast Exit Point Calculations‘Walk-on-Spheres’ AlgorithmWalk-in-SubdomainsMonte Carlo Treatment of Boundary Conditions
Monte Carlo EstimatesMonte Carlo EstimatesComputational GeometryCorrelated and Uncorrelated Sampling
Computational ResultsConclusions and Future Work
![Page 4: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/4.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
IntroductionA Little History on Monte Carlo Methods for PDEsSome Examples Using This for Computing Elliptic Problems
The Walk on Spheres MethodProblems in Electrostatics/MaterialsVarious Acceleration Techniques for Elliptic PDEs
Mathematical ModelElectrostatic Potential and EnergyThe Feynman-Kac Formula
Fast Exit Point Calculations‘Walk-on-Spheres’ AlgorithmWalk-in-SubdomainsMonte Carlo Treatment of Boundary Conditions
Monte Carlo EstimatesMonte Carlo EstimatesComputational GeometryCorrelated and Uncorrelated Sampling
Computational ResultsConclusions and Future Work
![Page 5: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/5.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
IntroductionA Little History on Monte Carlo Methods for PDEsSome Examples Using This for Computing Elliptic Problems
The Walk on Spheres MethodProblems in Electrostatics/MaterialsVarious Acceleration Techniques for Elliptic PDEs
Mathematical ModelElectrostatic Potential and EnergyThe Feynman-Kac Formula
Fast Exit Point Calculations‘Walk-on-Spheres’ AlgorithmWalk-in-SubdomainsMonte Carlo Treatment of Boundary Conditions
Monte Carlo EstimatesMonte Carlo EstimatesComputational GeometryCorrelated and Uncorrelated Sampling
Computational ResultsConclusions and Future Work
![Page 6: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/6.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
IntroductionA Little History on Monte Carlo Methods for PDEsSome Examples Using This for Computing Elliptic Problems
The Walk on Spheres MethodProblems in Electrostatics/MaterialsVarious Acceleration Techniques for Elliptic PDEs
Mathematical ModelElectrostatic Potential and EnergyThe Feynman-Kac Formula
Fast Exit Point Calculations‘Walk-on-Spheres’ AlgorithmWalk-in-SubdomainsMonte Carlo Treatment of Boundary Conditions
Monte Carlo EstimatesMonte Carlo EstimatesComputational GeometryCorrelated and Uncorrelated Sampling
Computational ResultsConclusions and Future Work
![Page 7: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/7.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
IntroductionA Little History on Monte Carlo Methods for PDEsSome Examples Using This for Computing Elliptic Problems
The Walk on Spheres MethodProblems in Electrostatics/MaterialsVarious Acceleration Techniques for Elliptic PDEs
Mathematical ModelElectrostatic Potential and EnergyThe Feynman-Kac Formula
Fast Exit Point Calculations‘Walk-on-Spheres’ AlgorithmWalk-in-SubdomainsMonte Carlo Treatment of Boundary Conditions
Monte Carlo EstimatesMonte Carlo EstimatesComputational GeometryCorrelated and Uncorrelated Sampling
Computational ResultsConclusions and Future Work
![Page 8: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/8.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
IntroductionA Little History on Monte Carlo Methods for PDEsSome Examples Using This for Computing Elliptic Problems
The Walk on Spheres MethodProblems in Electrostatics/MaterialsVarious Acceleration Techniques for Elliptic PDEs
Mathematical ModelElectrostatic Potential and EnergyThe Feynman-Kac Formula
Fast Exit Point Calculations‘Walk-on-Spheres’ AlgorithmWalk-in-SubdomainsMonte Carlo Treatment of Boundary Conditions
Monte Carlo EstimatesMonte Carlo EstimatesComputational GeometryCorrelated and Uncorrelated Sampling
Computational ResultsConclusions and Future Work
![Page 9: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/9.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
A Little History on Monte Carlo Methods for PDEs
Early History of MCMs for PDEs
1. Courant, Friedrichs, and Lewy: Their pivotal 1928 paper hasprobabilistic interpretations and MC algorithms for linear ellipticand parabolic problems
2. Fermi/Ulam/von Neumann: Atomic bomb calculations were doneusing Monte Carlo methods for neutron transport, their successinspired much post-War work especially in nuclear reactor design
3. Kac and Donsker: Used large deviation calculations to estimateeigenvalues of a linear Schrödinger equation
4. Forsythe and Leibler: Derived a MCM for solving special linearsystems related to discrete elliptic PDE problems
![Page 10: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/10.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
A Little History on Monte Carlo Methods for PDEs
Early History of MCMs for PDEs
1. Courant, Friedrichs, and Lewy: Their pivotal 1928 paper hasprobabilistic interpretations and MC algorithms for linear ellipticand parabolic problems
2. Fermi/Ulam/von Neumann: Atomic bomb calculations were doneusing Monte Carlo methods for neutron transport, their successinspired much post-War work especially in nuclear reactor design
3. Kac and Donsker: Used large deviation calculations to estimateeigenvalues of a linear Schrödinger equation
4. Forsythe and Leibler: Derived a MCM for solving special linearsystems related to discrete elliptic PDE problems
![Page 11: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/11.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
A Little History on Monte Carlo Methods for PDEs
Early History of MCMs for PDEs
1. Courant, Friedrichs, and Lewy: Their pivotal 1928 paper hasprobabilistic interpretations and MC algorithms for linear ellipticand parabolic problems
2. Fermi/Ulam/von Neumann: Atomic bomb calculations were doneusing Monte Carlo methods for neutron transport, their successinspired much post-War work especially in nuclear reactor design
3. Kac and Donsker: Used large deviation calculations to estimateeigenvalues of a linear Schrödinger equation
4. Forsythe and Leibler: Derived a MCM for solving special linearsystems related to discrete elliptic PDE problems
![Page 12: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/12.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
A Little History on Monte Carlo Methods for PDEs
Early History of MCMs for PDEs
1. Courant, Friedrichs, and Lewy: Their pivotal 1928 paper hasprobabilistic interpretations and MC algorithms for linear ellipticand parabolic problems
2. Fermi/Ulam/von Neumann: Atomic bomb calculations were doneusing Monte Carlo methods for neutron transport, their successinspired much post-War work especially in nuclear reactor design
3. Kac and Donsker: Used large deviation calculations to estimateeigenvalues of a linear Schrödinger equation
4. Forsythe and Leibler: Derived a MCM for solving special linearsystems related to discrete elliptic PDE problems
![Page 13: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/13.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
The Walk on Spheres Method
The First Passage (FP) Probability is the Green’sFunction
Back to our canonical elliptic boundary value problem:
12
∆u(x) = 0, x ∈ Ω
u(x) = f (x), x ∈ ∂Ω
I Distribution of z is uniform on the sphereI Mean of the values of u(z) over the sphere is u(x)
I u(x) has mean-value property and harmonicI Also, u(x) satisfies the boundary condition
u(x) = Ex [f (X x(t∂Ω))] (1)
![Page 14: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/14.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
The Walk on Spheres Method
The First Passage (FP) Probability is the Green’sFunction
Back to our canonical elliptic boundary value problem:
12
∆u(x) = 0, x ∈ Ω
u(x) = f (x), x ∈ ∂Ω
I Distribution of z is uniform on the sphereI Mean of the values of u(z) over the sphere is u(x)
I u(x) has mean-value property and harmonicI Also, u(x) satisfies the boundary condition
u(x) = Ex [f (X x(t∂Ω))] (1)
![Page 15: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/15.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
The Walk on Spheres Method
The First Passage (FP) Probability is the Green’sFunction
Back to our canonical elliptic boundary value problem:
12
∆u(x) = 0, x ∈ Ω
u(x) = f (x), x ∈ ∂Ω
I Distribution of z is uniform on the sphereI Mean of the values of u(z) over the sphere is u(x)
I u(x) has mean-value property and harmonicI Also, u(x) satisfies the boundary condition
u(x) = Ex [f (X x(t∂Ω))] (1)
![Page 16: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/16.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
The Walk on Spheres Method
The First Passage (FP) Probability is the Green’sFunction
Back to our canonical elliptic boundary value problem:
12
∆u(x) = 0, x ∈ Ω
u(x) = f (x), x ∈ ∂Ω
I Distribution of z is uniform on the sphereI Mean of the values of u(z) over the sphere is u(x)
I u(x) has mean-value property and harmonicI Also, u(x) satisfies the boundary condition
u(x) = Ex [f (X x(t∂Ω))] (1)
![Page 17: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/17.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
The Walk on Spheres Method
The First Passage (FP) Probability is the Green’sFunction
Reinterpreting as an average of the boundary values
u(x) =
∫∂Ω
p(x , y) f (y) dy (2)
Another representation in terms of an integral over the boundary
u(x) =
∫∂Ω
∂g(x , y)
∂nf (y) dy (3)
g(x , y) – Green’s function of the Dirichlet problem in Ω
=⇒ p(x , y) =∂g(x , y)
∂n(4)
![Page 18: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/18.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
The Walk on Spheres Method
‘Walk on Spheres’ (WOS) and ‘Green’s Function FirstPassage’ (GFFP) Algorithms
I Green’s function is known=⇒ direct simulation of exit points and computation of thesolution through averaging boundary values
I Green’s function is unknown=⇒ simulation of exit points from standard subdomains of Ω,e.g. spheres=⇒ Markov chain of ‘Walk on Spheres’ (or GFFP algorithm)x0 = x , x1, . . . , xNxi → ∂Ω and hits ε-shell is N = O(| ln(ε)|) stepsxN simulates exit point from Ω with O(ε) accuracy
![Page 19: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/19.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
The Walk on Spheres Method
‘Walk on Spheres’ (WOS) and ‘Green’s Function FirstPassage’ (GFFP) Algorithms
I Green’s function is known=⇒ direct simulation of exit points and computation of thesolution through averaging boundary values
I Green’s function is unknown=⇒ simulation of exit points from standard subdomains of Ω,e.g. spheres=⇒ Markov chain of ‘Walk on Spheres’ (or GFFP algorithm)x0 = x , x1, . . . , xNxi → ∂Ω and hits ε-shell is N = O(| ln(ε)|) stepsxN simulates exit point from Ω with O(ε) accuracy
![Page 20: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/20.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
The Walk on Spheres Method
‘Walk on Spheres’ (WOS) and ‘Green’s Function FirstPassage’ (GFFP) Algorithms
WOS:
![Page 21: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/21.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
The Walk on Spheres Method
Timing with WOS
![Page 22: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/22.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
Problems in Electrostatics/Materials
Solc-Stockmayer Model without Potential
![Page 23: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/23.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
Problems in Electrostatics/Materials
The Simulation-Tabulation (S-T) Method forGeneralization
I Green’s function for the non-intersected surface of a spherelocated on the surface of a reflecting sphere
![Page 24: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/24.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
Problems in Electrostatics/Materials
Porous Media: Complicated Interfaces
![Page 25: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/25.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
Problems in Electrostatics/Materials
Computing Capacitance Probabilistically
I Hubbard-Douglas: can compute permeability of nonskew objectvia capacitance
I Recall that C = Qu , if we hold conductor (Ω)at unit potential
u = 1, then C = total charge on conductor (surface)I The PDE system for the potential is
∆u = 0, x /∈ Ω; u = 1, x ∈ ∂Ω; u → 0 as x →∞ (5)
I Recall u(x) = Ex [f (X x(t∂Ω))] = probability of walker starting at xhitting Ω before escaping to infinity
I Charge density is first passage probabilityI Capacitance (relative to a sphere) is probability of walker starting
at x (random chosen on sphere) hitting Ω before escaping toinfinity
![Page 26: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/26.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
Problems in Electrostatics/Materials
Computing Capacitance Probabilistically
I Hubbard-Douglas: can compute permeability of nonskew objectvia capacitance
I Recall that C = Qu , if we hold conductor (Ω)at unit potential
u = 1, then C = total charge on conductor (surface)I The PDE system for the potential is
∆u = 0, x /∈ Ω; u = 1, x ∈ ∂Ω; u → 0 as x →∞ (5)
I Recall u(x) = Ex [f (X x(t∂Ω))] = probability of walker starting at xhitting Ω before escaping to infinity
I Charge density is first passage probabilityI Capacitance (relative to a sphere) is probability of walker starting
at x (random chosen on sphere) hitting Ω before escaping toinfinity
![Page 27: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/27.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
Problems in Electrostatics/Materials
Computing Capacitance Probabilistically
I Hubbard-Douglas: can compute permeability of nonskew objectvia capacitance
I Recall that C = Qu , if we hold conductor (Ω)at unit potential
u = 1, then C = total charge on conductor (surface)I The PDE system for the potential is
∆u = 0, x /∈ Ω; u = 1, x ∈ ∂Ω; u → 0 as x →∞ (5)
I Recall u(x) = Ex [f (X x(t∂Ω))] = probability of walker starting at xhitting Ω before escaping to infinity
I Charge density is first passage probabilityI Capacitance (relative to a sphere) is probability of walker starting
at x (random chosen on sphere) hitting Ω before escaping toinfinity
![Page 28: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/28.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
Problems in Electrostatics/Materials
Computing Capacitance Probabilistically
I Hubbard-Douglas: can compute permeability of nonskew objectvia capacitance
I Recall that C = Qu , if we hold conductor (Ω)at unit potential
u = 1, then C = total charge on conductor (surface)I The PDE system for the potential is
∆u = 0, x /∈ Ω; u = 1, x ∈ ∂Ω; u → 0 as x →∞ (5)
I Recall u(x) = Ex [f (X x(t∂Ω))] = probability of walker starting at xhitting Ω before escaping to infinity
I Charge density is first passage probabilityI Capacitance (relative to a sphere) is probability of walker starting
at x (random chosen on sphere) hitting Ω before escaping toinfinity
![Page 29: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/29.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
Problems in Electrostatics/Materials
Computing Capacitance Probabilistically
I Hubbard-Douglas: can compute permeability of nonskew objectvia capacitance
I Recall that C = Qu , if we hold conductor (Ω)at unit potential
u = 1, then C = total charge on conductor (surface)I The PDE system for the potential is
∆u = 0, x /∈ Ω; u = 1, x ∈ ∂Ω; u → 0 as x →∞ (5)
I Recall u(x) = Ex [f (X x(t∂Ω))] = probability of walker starting at xhitting Ω before escaping to infinity
I Charge density is first passage probabilityI Capacitance (relative to a sphere) is probability of walker starting
at x (random chosen on sphere) hitting Ω before escaping toinfinity
![Page 30: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/30.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
Problems in Electrostatics/Materials
Computing Capacitance Probabilistically
I Hubbard-Douglas: can compute permeability of nonskew objectvia capacitance
I Recall that C = Qu , if we hold conductor (Ω)at unit potential
u = 1, then C = total charge on conductor (surface)I The PDE system for the potential is
∆u = 0, x /∈ Ω; u = 1, x ∈ ∂Ω; u → 0 as x →∞ (5)
I Recall u(x) = Ex [f (X x(t∂Ω))] = probability of walker starting at xhitting Ω before escaping to infinity
I Charge density is first passage probabilityI Capacitance (relative to a sphere) is probability of walker starting
at x (random chosen on sphere) hitting Ω before escaping toinfinity
![Page 31: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/31.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
Problems in Electrostatics/Materials
Various Laplacian Green’s Functions for Green’sFunction First Passage (GFFP)
OO
O
Putting back (a) Void space(b) Intersecting(c)
![Page 32: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/32.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
Problems in Electrostatics/Materials
Escape to∞ in A Single Step
I Probability that a diffusing particle at r0 > b will escape to infinity
Pesc = 1− br0
= 1− α (6)
I Putting-back distribution density function
ω(θ, φ) =1− α2
4π[1− 2α cos θ + α2]3/2 (7)
I (b, θ, φ) ; spherical coordinates of the new position when the oldposition is put on the polar axis
![Page 33: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/33.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
Problems in Electrostatics/Materials
Escape to∞ in A Single Step
I Probability that a diffusing particle at r0 > b will escape to infinity
Pesc = 1− br0
= 1− α (6)
I Putting-back distribution density function
ω(θ, φ) =1− α2
4π[1− 2α cos θ + α2]3/2 (7)
I (b, θ, φ) ; spherical coordinates of the new position when the oldposition is put on the polar axis
![Page 34: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/34.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
Problems in Electrostatics/Materials
Escape to∞ in A Single Step
I Probability that a diffusing particle at r0 > b will escape to infinity
Pesc = 1− br0
= 1− α (6)
I Putting-back distribution density function
ω(θ, φ) =1− α2
4π[1− 2α cos θ + α2]3/2 (7)
I (b, θ, φ) ; spherical coordinates of the new position when the oldposition is put on the polar axis
![Page 35: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/35.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
Problems in Electrostatics/Materials
Charge Density on a Circular Disk via Last-Passage
![Page 36: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/36.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
Problems in Electrostatics/Materials
Time Reversal Brownian Motion: Approach from theOutside
![Page 37: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/37.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
Problems in Electrostatics/Materials
Approach from the Outside
I P(x): prob. of diffusing from ε above lower FP surface to∞
P(x) =
∫∂Ωy
g(x , y , ε)p(y ,∞)dS (8)
σ(x) = − 14π
ddε
∣∣∣∣∣ε=0
φ(x) =1
4πddε
∣∣∣∣∣ε=0
P(x) (9)
σ(x) =1
4π
∫∂Ωy
G(x , y)p(y ,∞)dS (10)
where
G(x , y) =ddε
∣∣∣∣∣ε=0
g(x , y , ε) (11)
I G(x , y) satisfies a point dipole problem
![Page 38: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/38.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
Problems in Electrostatics/Materials
Approach from the Outside
I P(x): prob. of diffusing from ε above lower FP surface to∞
P(x) =
∫∂Ωy
g(x , y , ε)p(y ,∞)dS (8)
σ(x) = − 14π
ddε
∣∣∣∣∣ε=0
φ(x) =1
4πddε
∣∣∣∣∣ε=0
P(x) (9)
σ(x) =1
4π
∫∂Ωy
G(x , y)p(y ,∞)dS (10)
where
G(x , y) =ddε
∣∣∣∣∣ε=0
g(x , y , ε) (11)
I G(x , y) satisfies a point dipole problem
![Page 39: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/39.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
Problems in Electrostatics/Materials
Charge Density on the Circular Disk
G =34
cos θa3 (12)
σ(x) =3
16π
∫∂Ωr
cos θa3 p(r,∞)dS (13)
wherep(r,∞) = 1− 2
πarctan
√2b√
r2 − b2 +√
(r2 − b2)2 + 4b2x2
(14)
![Page 40: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/40.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
Problems in Electrostatics/Materials
Charge Density on the Circular Disk
![Page 41: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/41.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
Problems in Electrostatics/Materials
Edge Distribution on the Circular Disk
σ(r) =1
4π1√
1− r2(15)
Let r = 1− x :σ(x) =
14π
1√2x
(1− x/2)−1/2 (16)
when x is small enough,σ(x) ' 1
4√
2π1√x
(17)
σ(x) ' σe1√x
(18)
![Page 42: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/42.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
Problems in Electrostatics/Materials
Unit Cube Edge Distribution
![Page 43: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/43.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
Problems in Electrostatics/Materials
Unit Cube Edge Distribution
σ(x , δe) = δπ/α−1e σe(x) (19)
I σ(x , δe): charge on a curve parallel to the edge separated by δe
I σe(x): edge distributionI α: angle between the two intersecting surfaces, here α = 3π/2
σe(x) =1
4πlimδe→0
δ1−π/αe
∫∂Ωe
G(x , y)p(y ,∞)dS (20)
I ∂Ωe: cylindrical surface that intersects the pair of absorbingsurfaces meeting at angle α
![Page 44: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/44.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
Problems in Electrostatics/Materials
Unit Cube Edge Distribution
σ(x , δe) = δπ/α−1e σe(x) (19)
I σ(x , δe): charge on a curve parallel to the edge separated by δe
I σe(x): edge distributionI α: angle between the two intersecting surfaces, here α = 3π/2
σe(x) =1
4πlimδe→0
δ1−π/αe
∫∂Ωe
G(x , y)p(y ,∞)dS (20)
I ∂Ωe: cylindrical surface that intersects the pair of absorbingsurfaces meeting at angle α
![Page 45: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/45.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
Problems in Electrostatics/Materials
Unit Cube Edge Distribution
σ(x , δe) = δπ/α−1e σe(x) (19)
I σ(x , δe): charge on a curve parallel to the edge separated by δe
I σe(x): edge distributionI α: angle between the two intersecting surfaces, here α = 3π/2
σe(x) =1
4πlimδe→0
δ1−π/αe
∫∂Ωe
G(x , y)p(y ,∞)dS (20)
I ∂Ωe: cylindrical surface that intersects the pair of absorbingsurfaces meeting at angle α
![Page 46: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/46.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
Problems in Electrostatics/Materials
Unit Cube Edge Distribution
σ(x , δe) = δπ/α−1e σe(x) (19)
I σ(x , δe): charge on a curve parallel to the edge separated by δe
I σe(x): edge distributionI α: angle between the two intersecting surfaces, here α = 3π/2
σe(x) =1
4πlimδe→0
δ1−π/αe
∫∂Ωe
G(x , y)p(y ,∞)dS (20)
I ∂Ωe: cylindrical surface that intersects the pair of absorbingsurfaces meeting at angle α
![Page 47: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/47.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
Problems in Electrostatics/Materials
Unit Cube Edge Distribution
I G(x , y):
G(x , y) =d
dδε
∣∣∣∣∣δε=0
g(x , y , δε) (21)
I g(x , y , δε): Laplace Green’s function on the surface, ∂Ωe, withsource point x at a distance δε from the absorbing surface
I p(y ,∞): probability that a diffusing particle, initiated at pointy ∈ ∂Ωe, diffuses to infinity without returning to the absorbingsurface
![Page 48: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/48.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
Problems in Electrostatics/Materials
Unit Cube Edge Distribution
I G(x , y):
G(x , y) =d
dδε
∣∣∣∣∣δε=0
g(x , y , δε) (21)
I g(x , y , δε): Laplace Green’s function on the surface, ∂Ωe, withsource point x at a distance δε from the absorbing surface
I p(y ,∞): probability that a diffusing particle, initiated at pointy ∈ ∂Ωe, diffuses to infinity without returning to the absorbingsurface
![Page 49: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/49.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
Problems in Electrostatics/Materials
Unit Cube Edge Distribution
I G(x , y):
G(x , y) =d
dδε
∣∣∣∣∣δε=0
g(x , y , δε) (21)
I g(x , y , δε): Laplace Green’s function on the surface, ∂Ωe, withsource point x at a distance δε from the absorbing surface
I p(y ,∞): probability that a diffusing particle, initiated at pointy ∈ ∂Ωe, diffuses to infinity without returning to the absorbingsurface
![Page 50: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/50.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
Problems in Electrostatics/Materials
Unit Cube Edge Distribution
G(ρ = a, φ, z) =1
Γ(5/3)22/3
49πLa
∞∑n=1
sin(2
3φ)
sin(nπz
L
)sin(nπz ′
L
)×(nπ
L
)2/3 1I2/3( nπa
L )
G(ρ, φ, z = 0) =1
Γ(5/3)22/3
49πL
∞∑n=1
sin(2
3φ)(nπ
L
)5/3sin(nπz ′
L
)× 1
I2/3( nπaL )
[I2/3
(nπaL
)K2/3
(nπρL
)− K2/3
(nπaL
)I2/3
(nπρL
)]
![Page 51: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/51.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
Problems in Electrostatics/Materials
Unit Cube Edge Distribution
Figure: First- and last-passage edge computations
![Page 52: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/52.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
Problems in Electrostatics/Materials
Unit Cube Edge Distribution
Figure: The slope, that is, the exponent of the edge distribution near thecorner is approximately −0.20, that is, σe ∼ δ−1/5
c
![Page 53: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/53.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
Various Acceleration Techniques for Elliptic PDEs
Walk on the Boundary Algorithm
I µ(y) = − 14π
∂φ
∂n(y) ; surface charge density
I φ(x) =
∫∂Ω
1|x − y |
µ(y)dσ(y) ; electrostatic potential
Limit properties of the normal derivative (x → y outside of Ω):
µ(y) =
∫∂Ω
n(y) · (y − y ′)2π|y − y ′|3
µ(y ′)dσ(y ′)
By the ergodic theorem (convex Ω)∫∂Ω
v(y)π∞(y)dσ(y) = limN→∞
1N
N∑n=1
v(yn)
![Page 54: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/54.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
Various Acceleration Techniques for Elliptic PDEs
Walk on the Boundary Algorithm
I µ(y) = − 14π
∂φ
∂n(y) ; surface charge density
I φ(x) =
∫∂Ω
1|x − y |
µ(y)dσ(y) ; electrostatic potential
Limit properties of the normal derivative (x → y outside of Ω):
µ(y) =
∫∂Ω
n(y) · (y − y ′)2π|y − y ′|3
µ(y ′)dσ(y ′)
By the ergodic theorem (convex Ω)∫∂Ω
v(y)π∞(y)dσ(y) = limN→∞
1N
N∑n=1
v(yn)
![Page 55: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/55.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
Various Acceleration Techniques for Elliptic PDEs
Walk on the Boundary Algorithm
I π∞ - stationary distribution of Markov chain yn with transition
density p(yn → yn+1) =n(yn+1) · (yn+1 − yn)
2π|yn+1 − yn|3I µ = Cπ∞I C - capacitance if φ|∂Ω = 1I φ(x) = 1 for x ∈ Ω
C = ( limN→∞
1N
N∑n=1
v(yn))−1 for v(y) =1
x − y
![Page 56: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/56.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
Various Acceleration Techniques for Elliptic PDEs
Walk on the Boundary Algorithm
I π∞ - stationary distribution of Markov chain yn with transition
density p(yn → yn+1) =n(yn+1) · (yn+1 − yn)
2π|yn+1 − yn|3I µ = Cπ∞I C - capacitance if φ|∂Ω = 1I φ(x) = 1 for x ∈ Ω
C = ( limN→∞
1N
N∑n=1
v(yn))−1 for v(y) =1
x − y
![Page 57: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/57.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
Various Acceleration Techniques for Elliptic PDEs
Walk on the Boundary Algorithm
I π∞ - stationary distribution of Markov chain yn with transition
density p(yn → yn+1) =n(yn+1) · (yn+1 − yn)
2π|yn+1 − yn|3I µ = Cπ∞I C - capacitance if φ|∂Ω = 1I φ(x) = 1 for x ∈ Ω
C = ( limN→∞
1N
N∑n=1
v(yn))−1 for v(y) =1
x − y
![Page 58: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/58.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
Various Acceleration Techniques for Elliptic PDEs
Walk on the Boundary Algorithm
I π∞ - stationary distribution of Markov chain yn with transition
density p(yn → yn+1) =n(yn+1) · (yn+1 − yn)
2π|yn+1 − yn|3I µ = Cπ∞I C - capacitance if φ|∂Ω = 1I φ(x) = 1 for x ∈ Ω
C = ( limN→∞
1N
N∑n=1
v(yn))−1 for v(y) =1
x − y
![Page 59: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/59.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
Various Acceleration Techniques for Elliptic PDEs
Capacitance of the Unit Cube
Reitan-Higgins (1951) 0.6555Greenspan-Silverman (1965) 0.661
Cochran (1967) 0.6596Goto-Shi-Yoshida (1992) 0.6615897 ± 5 × 10−7
Conjectured Hubbard-Douglas (1993) 0.65946...Douglas-Zhou-Hubbard (1994) 0.6632 ± 0.0003Given-Hubbard-Douglas (1997) 0.660675 ± 0.00001
Read (1997) 0.6606785± 0.000003First passage method (2001) 0.660683± 0.000005
Walk on boundary algorithm (2002) 0.6606780± 0.0000004
![Page 60: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/60.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
Various Acceleration Techniques for Elliptic PDEs
Continuum Biochemical ElectrostaticsMotivation
I Experimental Data: Folding, stability & binding behavior ofbiomolecules can be modulated by changes in salt concentration
I Physical Model: Implicit solvent-based Poisson-Boltzmann modelcan provide accurate predictions of salt dependent behavior ofbiomolecules
I Mathematical Model: Elliptic boundary-value problemsSpecific Problems
I Electrostatic free energy for linear case: only finite number ofelectrostatic potential point values
I Dependence of energy on geometry: needs accurate treatmentI Singularities in solution: have to be taken into account
analyticallyI Behavior at infinity: must be exactly enforcedI Functional dependence on salt concentration: needs accurate
estimate
![Page 61: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/61.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
Various Acceleration Techniques for Elliptic PDEs
Continuum Biochemical ElectrostaticsMotivation
I Experimental Data: Folding, stability & binding behavior ofbiomolecules can be modulated by changes in salt concentration
I Physical Model: Implicit solvent-based Poisson-Boltzmann modelcan provide accurate predictions of salt dependent behavior ofbiomolecules
I Mathematical Model: Elliptic boundary-value problemsSpecific Problems
I Electrostatic free energy for linear case: only finite number ofelectrostatic potential point values
I Dependence of energy on geometry: needs accurate treatmentI Singularities in solution: have to be taken into account
analyticallyI Behavior at infinity: must be exactly enforcedI Functional dependence on salt concentration: needs accurate
estimate
![Page 62: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/62.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
Various Acceleration Techniques for Elliptic PDEs
Continuum Biochemical ElectrostaticsMotivation
I Experimental Data: Folding, stability & binding behavior ofbiomolecules can be modulated by changes in salt concentration
I Physical Model: Implicit solvent-based Poisson-Boltzmann modelcan provide accurate predictions of salt dependent behavior ofbiomolecules
I Mathematical Model: Elliptic boundary-value problemsSpecific Problems
I Electrostatic free energy for linear case: only finite number ofelectrostatic potential point values
I Dependence of energy on geometry: needs accurate treatmentI Singularities in solution: have to be taken into account
analyticallyI Behavior at infinity: must be exactly enforcedI Functional dependence on salt concentration: needs accurate
estimate
![Page 63: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/63.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
Various Acceleration Techniques for Elliptic PDEs
Continuum Biochemical ElectrostaticsMotivation
I Experimental Data: Folding, stability & binding behavior ofbiomolecules can be modulated by changes in salt concentration
I Physical Model: Implicit solvent-based Poisson-Boltzmann modelcan provide accurate predictions of salt dependent behavior ofbiomolecules
I Mathematical Model: Elliptic boundary-value problemsSpecific Problems
I Electrostatic free energy for linear case: only finite number ofelectrostatic potential point values
I Dependence of energy on geometry: needs accurate treatmentI Singularities in solution: have to be taken into account
analyticallyI Behavior at infinity: must be exactly enforcedI Functional dependence on salt concentration: needs accurate
estimate
![Page 64: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/64.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
Various Acceleration Techniques for Elliptic PDEs
Continuum Biochemical ElectrostaticsMotivation
I Experimental Data: Folding, stability & binding behavior ofbiomolecules can be modulated by changes in salt concentration
I Physical Model: Implicit solvent-based Poisson-Boltzmann modelcan provide accurate predictions of salt dependent behavior ofbiomolecules
I Mathematical Model: Elliptic boundary-value problemsSpecific Problems
I Electrostatic free energy for linear case: only finite number ofelectrostatic potential point values
I Dependence of energy on geometry: needs accurate treatmentI Singularities in solution: have to be taken into account
analyticallyI Behavior at infinity: must be exactly enforcedI Functional dependence on salt concentration: needs accurate
estimate
![Page 65: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/65.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
Various Acceleration Techniques for Elliptic PDEs
Continuum Biochemical ElectrostaticsMotivation
I Experimental Data: Folding, stability & binding behavior ofbiomolecules can be modulated by changes in salt concentration
I Physical Model: Implicit solvent-based Poisson-Boltzmann modelcan provide accurate predictions of salt dependent behavior ofbiomolecules
I Mathematical Model: Elliptic boundary-value problemsSpecific Problems
I Electrostatic free energy for linear case: only finite number ofelectrostatic potential point values
I Dependence of energy on geometry: needs accurate treatmentI Singularities in solution: have to be taken into account
analyticallyI Behavior at infinity: must be exactly enforcedI Functional dependence on salt concentration: needs accurate
estimate
![Page 66: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/66.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
Various Acceleration Techniques for Elliptic PDEs
Continuum Biochemical ElectrostaticsMotivation
I Experimental Data: Folding, stability & binding behavior ofbiomolecules can be modulated by changes in salt concentration
I Physical Model: Implicit solvent-based Poisson-Boltzmann modelcan provide accurate predictions of salt dependent behavior ofbiomolecules
I Mathematical Model: Elliptic boundary-value problemsSpecific Problems
I Electrostatic free energy for linear case: only finite number ofelectrostatic potential point values
I Dependence of energy on geometry: needs accurate treatmentI Singularities in solution: have to be taken into account
analyticallyI Behavior at infinity: must be exactly enforcedI Functional dependence on salt concentration: needs accurate
estimate
![Page 67: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/67.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
Various Acceleration Techniques for Elliptic PDEs
Continuum Biochemical ElectrostaticsMotivation
I Experimental Data: Folding, stability & binding behavior ofbiomolecules can be modulated by changes in salt concentration
I Physical Model: Implicit solvent-based Poisson-Boltzmann modelcan provide accurate predictions of salt dependent behavior ofbiomolecules
I Mathematical Model: Elliptic boundary-value problemsSpecific Problems
I Electrostatic free energy for linear case: only finite number ofelectrostatic potential point values
I Dependence of energy on geometry: needs accurate treatmentI Singularities in solution: have to be taken into account
analyticallyI Behavior at infinity: must be exactly enforcedI Functional dependence on salt concentration: needs accurate
estimate
![Page 68: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/68.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
Various Acceleration Techniques for Elliptic PDEs
Monte Carlo Methods: PropertiesI Monte Carlo methods for solving Poisson and linearized
Poisson-Boltzmann equations (PBEs)I Analytical treatment of geometry, singularities, behavior at infinityI Capability to compute point values of solution (energies) and its
spatial derivatives (forces)I New methods for the flux boundary conditions (exact integral
formulation)I Simultaneous correlated computation of values at different salt
concentrations
![Page 69: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/69.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
Various Acceleration Techniques for Elliptic PDEs
Monte Carlo Methods: PropertiesI Monte Carlo methods for solving Poisson and linearized
Poisson-Boltzmann equations (PBEs)I Analytical treatment of geometry, singularities, behavior at infinityI Capability to compute point values of solution (energies) and its
spatial derivatives (forces)I New methods for the flux boundary conditions (exact integral
formulation)I Simultaneous correlated computation of values at different salt
concentrations
![Page 70: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/70.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
Various Acceleration Techniques for Elliptic PDEs
Monte Carlo Methods: PropertiesI Monte Carlo methods for solving Poisson and linearized
Poisson-Boltzmann equations (PBEs)I Analytical treatment of geometry, singularities, behavior at infinityI Capability to compute point values of solution (energies) and its
spatial derivatives (forces)I New methods for the flux boundary conditions (exact integral
formulation)I Simultaneous correlated computation of values at different salt
concentrations
![Page 71: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/71.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
Various Acceleration Techniques for Elliptic PDEs
Monte Carlo Methods: PropertiesI Monte Carlo methods for solving Poisson and linearized
Poisson-Boltzmann equations (PBEs)I Analytical treatment of geometry, singularities, behavior at infinityI Capability to compute point values of solution (energies) and its
spatial derivatives (forces)I New methods for the flux boundary conditions (exact integral
formulation)I Simultaneous correlated computation of values at different salt
concentrations
![Page 72: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/72.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
Various Acceleration Techniques for Elliptic PDEs
Monte Carlo Methods: PropertiesI Monte Carlo methods for solving Poisson and linearized
Poisson-Boltzmann equations (PBEs)I Analytical treatment of geometry, singularities, behavior at infinityI Capability to compute point values of solution (energies) and its
spatial derivatives (forces)I New methods for the flux boundary conditions (exact integral
formulation)I Simultaneous correlated computation of values at different salt
concentrations
![Page 73: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/73.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Some Examples Using This for Computing Elliptic Problems
Various Acceleration Techniques for Elliptic PDEs
Mathematical Model: Molecular Geometry
Figure: Biomolecule with dielectric εi and region region Gi is in solution withdielectric εe and region Ge. On the boundary of the biomolecule, electrostaticpotential and normal component of dielectric displacement continue
![Page 74: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/74.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Mathematical Model
Mathematical Model: Partial Differential EquationsI Poisson equation for the electrostatic potential, Φi , and point
charges, Qm, inside a molecule (in CGS units):
εi ∆Φi (x) + 4πM∑
m=1
Qmδ(x − x (m)) = 0 , x ∈ Gi
I For 1-1 salt (such as NaCl) Poisson-Boltzmann equation (PBE):
∆Φe(x)− κ2 sinh(Φe(x)) = 0 , x ∈ Ge ,
but we only consider the linearized PBE:
∆Φe(x)− κ2Φe(x) = 0 , x ∈ Ge
I For one-surface model: continuity condition on the dielectricboundary
Φi = Φe , εi∂Φi
∂n(y)= εe
∂Φe
∂n(y), y ∈ Γ
![Page 75: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/75.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Mathematical Model
Mathematical Model: Partial Differential EquationsI Poisson equation for the electrostatic potential, Φi , and point
charges, Qm, inside a molecule (in CGS units):
εi ∆Φi (x) + 4πM∑
m=1
Qmδ(x − x (m)) = 0 , x ∈ Gi
I For 1-1 salt (such as NaCl) Poisson-Boltzmann equation (PBE):
∆Φe(x)− κ2 sinh(Φe(x)) = 0 , x ∈ Ge ,
but we only consider the linearized PBE:
∆Φe(x)− κ2Φe(x) = 0 , x ∈ Ge
I For one-surface model: continuity condition on the dielectricboundary
Φi = Φe , εi∂Φi
∂n(y)= εe
∂Φe
∂n(y), y ∈ Γ
![Page 76: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/76.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Mathematical Model
Mathematical Model: Partial Differential EquationsI Poisson equation for the electrostatic potential, Φi , and point
charges, Qm, inside a molecule (in CGS units):
εi ∆Φi (x) + 4πM∑
m=1
Qmδ(x − x (m)) = 0 , x ∈ Gi
I For 1-1 salt (such as NaCl) Poisson-Boltzmann equation (PBE):
∆Φe(x)− κ2 sinh(Φe(x)) = 0 , x ∈ Ge ,
but we only consider the linearized PBE:
∆Φe(x)− κ2Φe(x) = 0 , x ∈ Ge
I For one-surface model: continuity condition on the dielectricboundary
Φi = Φe , εi∂Φi
∂n(y)= εe
∂Φe
∂n(y), y ∈ Γ
![Page 77: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/77.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Mathematical Model
Mathematical Model: Debye-Hückle Parameter
Dependence on salt in the Debye-Hückle parameter (units as perKirkwood):
κ2 =8πNAe2Cs
εe1000kBT, where
I Cs – concentration of ions (in moles)I NA – Avogadro’s numberI e – elementary protonic chargeI kB – Boltzmann’s constantI εe – dielectric permittivity outside the molecule
![Page 78: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/78.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Mathematical Model
Mathematical Model: Debye-Hückle Parameter
Dependence on salt in the Debye-Hückle parameter (units as perKirkwood):
κ2 =8πNAe2Cs
εe1000kBT, where
I Cs – concentration of ions (in moles)I NA – Avogadro’s numberI e – elementary protonic chargeI kB – Boltzmann’s constantI εe – dielectric permittivity outside the molecule
![Page 79: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/79.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Mathematical Model
Mathematical Model: Debye-Hückle Parameter
Dependence on salt in the Debye-Hückle parameter (units as perKirkwood):
κ2 =8πNAe2Cs
εe1000kBT, where
I Cs – concentration of ions (in moles)I NA – Avogadro’s numberI e – elementary protonic chargeI kB – Boltzmann’s constantI εe – dielectric permittivity outside the molecule
![Page 80: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/80.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Mathematical Model
Mathematical Model: Debye-Hückle Parameter
Dependence on salt in the Debye-Hückle parameter (units as perKirkwood):
κ2 =8πNAe2Cs
εe1000kBT, where
I Cs – concentration of ions (in moles)I NA – Avogadro’s numberI e – elementary protonic chargeI kB – Boltzmann’s constantI εe – dielectric permittivity outside the molecule
![Page 81: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/81.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Mathematical Model
Mathematical Model: Debye-Hückle Parameter
Dependence on salt in the Debye-Hückle parameter (units as perKirkwood):
κ2 =8πNAe2Cs
εe1000kBT, where
I Cs – concentration of ions (in moles)I NA – Avogadro’s numberI e – elementary protonic chargeI kB – Boltzmann’s constantI εe – dielectric permittivity outside the molecule
![Page 82: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/82.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Mathematical Model
Electrostatic Potential and Energy
Electrostatic Potential and EnergyI Point values of the potential: Φ(x) = Φrf (x) + Φc(x)
Here, singular part of Φ:
Φc(x) =M∑
m=1
Qm
|x − x (m)|
I Reaction field electrostatic free energy of a molecule is linearcombination of point values of the regular part of the electrostaticpotential:
Wrf =12
M∑m=1
Φrf (x (m))Qm ,
I Electrostatic solvation free energy = difference between theenergy for a molecule in solvent with a given salt concentrationand the energy for the same molecule in vacuum:
∆Gelecsolv = Wrf (εi , εe, κ)−Wrf (εi ,1,0)
![Page 83: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/83.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Mathematical Model
Electrostatic Potential and Energy
Electrostatic Potential and EnergyI Point values of the potential: Φ(x) = Φrf (x) + Φc(x)
Here, singular part of Φ:
Φc(x) =M∑
m=1
Qm
|x − x (m)|
I Reaction field electrostatic free energy of a molecule is linearcombination of point values of the regular part of the electrostaticpotential:
Wrf =12
M∑m=1
Φrf (x (m))Qm ,
I Electrostatic solvation free energy = difference between theenergy for a molecule in solvent with a given salt concentrationand the energy for the same molecule in vacuum:
∆Gelecsolv = Wrf (εi , εe, κ)−Wrf (εi ,1,0)
![Page 84: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/84.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Mathematical Model
Electrostatic Potential and Energy
Electrostatic Potential and EnergyI Point values of the potential: Φ(x) = Φrf (x) + Φc(x)
Here, singular part of Φ:
Φc(x) =M∑
m=1
Qm
|x − x (m)|
I Reaction field electrostatic free energy of a molecule is linearcombination of point values of the regular part of the electrostaticpotential:
Wrf =12
M∑m=1
Φrf (x (m))Qm ,
I Electrostatic solvation free energy = difference between theenergy for a molecule in solvent with a given salt concentrationand the energy for the same molecule in vacuum:
∆Gelecsolv = Wrf (εi , εe, κ)−Wrf (εi ,1,0)
![Page 85: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/85.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Mathematical Model
The Feynman-Kac Formula
The Feynman-Kac Formula
I Consider the Dirichlet problem for the Poisson equation in thedomain Ω ∈ Rd
−12
∆u(x) = g(x), x ∈ Ω, u(x) = f (x), x ∈ ∂Ω
I If we assume g(x) = 0, then we have the Laplace equation, andthe solution at the point y ∈ Ω is given as the following Brownianmotion expectation:
u(y) = E[f (βy (τ∂Ω))],
where βy (·) is Brownian motion starting at the point y , and τ∂Ω isthe first-passage time of this Brownian motion,i.e. τ∂Ω = inftβy (t) ∈ ∂Ω
![Page 86: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/86.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Mathematical Model
The Feynman-Kac Formula
The Feynman-Kac Formula
I Consider the Dirichlet problem for the Poisson equation in thedomain Ω ∈ Rd
−12
∆u(x) = g(x), x ∈ Ω, u(x) = f (x), x ∈ ∂Ω
I If we assume g(x) = 0, then we have the Laplace equation, andthe solution at the point y ∈ Ω is given as the following Brownianmotion expectation:
u(y) = E[f (βy (τ∂Ω))],
where βy (·) is Brownian motion starting at the point y , and τ∂Ω isthe first-passage time of this Brownian motion,i.e. τ∂Ω = inftβy (t) ∈ ∂Ω
![Page 87: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/87.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Mathematical Model
The Feynman-Kac Formula
The Feynman-Kac FormulaI If we set f (x) = 0 and have g(x) 6= 0, the solution is
u(y) = E[ ∫ τ∂Ω
0g(βy (s)) ds
]I By linear superposition, the solution to Poisson equation is given
probabilistically as
u(y) = E[ ∫ τ∂Ω
0g(βy (s)) ds + f (βy (τ∂Ω))
]I The linearized Poisson-Boltzmann equation is given by
∆u(x)−κ2u(x) = 0, x ∈ Ω, u(x) = f (x), x ∈ ∂Ω, u → 0 as |x | → ∞
and has Wiener integral representation:
u(y) = E[f (βy (τ∂Ω))e−
∫ τ∂Ω0 κ2 ds
]
![Page 88: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/88.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Mathematical Model
The Feynman-Kac Formula
The Feynman-Kac FormulaI If we set f (x) = 0 and have g(x) 6= 0, the solution is
u(y) = E[ ∫ τ∂Ω
0g(βy (s)) ds
]I By linear superposition, the solution to Poisson equation is given
probabilistically as
u(y) = E[ ∫ τ∂Ω
0g(βy (s)) ds + f (βy (τ∂Ω))
]I The linearized Poisson-Boltzmann equation is given by
∆u(x)−κ2u(x) = 0, x ∈ Ω, u(x) = f (x), x ∈ ∂Ω, u → 0 as |x | → ∞
and has Wiener integral representation:
u(y) = E[f (βy (τ∂Ω))e−
∫ τ∂Ω0 κ2 ds
]
![Page 89: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/89.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Mathematical Model
The Feynman-Kac Formula
The Feynman-Kac FormulaI If we set f (x) = 0 and have g(x) 6= 0, the solution is
u(y) = E[ ∫ τ∂Ω
0g(βy (s)) ds
]I By linear superposition, the solution to Poisson equation is given
probabilistically as
u(y) = E[ ∫ τ∂Ω
0g(βy (s)) ds + f (βy (τ∂Ω))
]I The linearized Poisson-Boltzmann equation is given by
∆u(x)−κ2u(x) = 0, x ∈ Ω, u(x) = f (x), x ∈ ∂Ω, u → 0 as |x | → ∞
and has Wiener integral representation:
u(y) = E[f (βy (τ∂Ω))e−
∫ τ∂Ω0 κ2 ds
]
![Page 90: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/90.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Fast Exit Point Calculations
‘Walk-on-Spheres’ Algorithm
‘Walk-on-Spheres’ Algorithm
I Walk-on-spheres (WOS) algorithm for general domains with aregular boundary
I Define a Markov chain xi , i = 1,2, . . .I Set x0 = x (m) for some m, xi = xi−1 + diωi , i = 1,2, . . ., where
1. di = d(xi−1) is distance from xi−1 to Γ2. ωi is sequence of independent unit isotropic vectors3. xi is the exit point from the ball, B(xi−1, d(xi−1)), for a Brownian
motion starting at xi−1
I Outside the molecule, on every step, walk-on-spheres terminates
with probability 1− q(κ,di ), where q(κ,di ) =κdi
sinh(κdi )to deal
with LPBE
![Page 91: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/91.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Fast Exit Point Calculations
‘Walk-on-Spheres’ Algorithm
‘Walk-on-Spheres’ Algorithm
I Walk-on-spheres (WOS) algorithm for general domains with aregular boundary
I Define a Markov chain xi , i = 1,2, . . .I Set x0 = x (m) for some m, xi = xi−1 + diωi , i = 1,2, . . ., where
1. di = d(xi−1) is distance from xi−1 to Γ2. ωi is sequence of independent unit isotropic vectors3. xi is the exit point from the ball, B(xi−1, d(xi−1)), for a Brownian
motion starting at xi−1
I Outside the molecule, on every step, walk-on-spheres terminates
with probability 1− q(κ,di ), where q(κ,di ) =κdi
sinh(κdi )to deal
with LPBE
![Page 92: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/92.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Fast Exit Point Calculations
‘Walk-on-Spheres’ Algorithm
‘Walk-on-Spheres’ Algorithm
I Walk-on-spheres (WOS) algorithm for general domains with aregular boundary
I Define a Markov chain xi , i = 1,2, . . .I Set x0 = x (m) for some m, xi = xi−1 + diωi , i = 1,2, . . ., where
1. di = d(xi−1) is distance from xi−1 to Γ2. ωi is sequence of independent unit isotropic vectors3. xi is the exit point from the ball, B(xi−1, d(xi−1)), for a Brownian
motion starting at xi−1
I Outside the molecule, on every step, walk-on-spheres terminates
with probability 1− q(κ,di ), where q(κ,di ) =κdi
sinh(κdi )to deal
with LPBE
![Page 93: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/93.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Fast Exit Point Calculations
‘Walk-on-Spheres’ Algorithm
‘Walk-on-Spheres’ Algorithm
I Walk-on-spheres (WOS) algorithm for general domains with aregular boundary
I Define a Markov chain xi , i = 1,2, . . .I Set x0 = x (m) for some m, xi = xi−1 + diωi , i = 1,2, . . ., where
1. di = d(xi−1) is distance from xi−1 to Γ2. ωi is sequence of independent unit isotropic vectors3. xi is the exit point from the ball, B(xi−1, d(xi−1)), for a Brownian
motion starting at xi−1
I Outside the molecule, on every step, walk-on-spheres terminates
with probability 1− q(κ,di ), where q(κ,di ) =κdi
sinh(κdi )to deal
with LPBE
![Page 94: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/94.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Fast Exit Point Calculations
‘Walk-on-Spheres’ Algorithm
‘Walk-on-Spheres’ Algorithm
I Walk-on-spheres (WOS) algorithm for general domains with aregular boundary
I Define a Markov chain xi , i = 1,2, . . .I Set x0 = x (m) for some m, xi = xi−1 + diωi , i = 1,2, . . ., where
1. di = d(xi−1) is distance from xi−1 to Γ2. ωi is sequence of independent unit isotropic vectors3. xi is the exit point from the ball, B(xi−1, d(xi−1)), for a Brownian
motion starting at xi−1
I Outside the molecule, on every step, walk-on-spheres terminates
with probability 1− q(κ,di ), where q(κ,di ) =κdi
sinh(κdi )to deal
with LPBE
![Page 95: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/95.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Fast Exit Point Calculations
‘Walk-on-Spheres’ Algorithm
‘Walk-on-Spheres’ Algorithm
I Walk-on-spheres (WOS) algorithm for general domains with aregular boundary
I Define a Markov chain xi , i = 1,2, . . .I Set x0 = x (m) for some m, xi = xi−1 + diωi , i = 1,2, . . ., where
1. di = d(xi−1) is distance from xi−1 to Γ2. ωi is sequence of independent unit isotropic vectors3. xi is the exit point from the ball, B(xi−1, d(xi−1)), for a Brownian
motion starting at xi−1
I Outside the molecule, on every step, walk-on-spheres terminates
with probability 1− q(κ,di ), where q(κ,di ) =κdi
sinh(κdi )to deal
with LPBE
![Page 96: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/96.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Fast Exit Point Calculations
‘Walk-on-Spheres’ Algorithm
‘Walk-on-Spheres’ Algorithm
I Walk-on-spheres (WOS) algorithm for general domains with aregular boundary
I Define a Markov chain xi , i = 1,2, . . .I Set x0 = x (m) for some m, xi = xi−1 + diωi , i = 1,2, . . ., where
1. di = d(xi−1) is distance from xi−1 to Γ2. ωi is sequence of independent unit isotropic vectors3. xi is the exit point from the ball, B(xi−1, d(xi−1)), for a Brownian
motion starting at xi−1
I Outside the molecule, on every step, walk-on-spheres terminates
with probability 1− q(κ,di ), where q(κ,di ) =κdi
sinh(κdi )to deal
with LPBE
![Page 97: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/97.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Fast Exit Point Calculations
Walk-in-Subdomains
‘Walk-on-Spheres’ and ‘Walk-in-Subdomains’I For general domains, an efficient way to simulate exit points is a
combination of1. Inside the molecule: ‘walk-in-subdomains’2. Outside the molecule ‘walk-on-spheres’
I The whole domain, Gi , is represented as a union of intersectingsubdomains:
Gi =M⋃
m=1
Gm
I ‘Walk-in-Subdomains’: Simulate exit point separately in every Gm
1. x0 = x , x1, . . . , xN – Markov chain, every xi+1 is an exit point fromthe corresponding subdomain for Brownian motion starting at xi
2. For spherical subdomains, B(xmi ,R
mi ), exit points are distributed in
accordance with the Poisson kernel:
14πRm
i
|xi − xmi |2 − (Rm
i )2
|xi − xi+1|3
![Page 98: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/98.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Fast Exit Point Calculations
Walk-in-Subdomains
‘Walk-on-Spheres’ and ‘Walk-in-Subdomains’I For general domains, an efficient way to simulate exit points is a
combination of1. Inside the molecule: ‘walk-in-subdomains’2. Outside the molecule ‘walk-on-spheres’
I The whole domain, Gi , is represented as a union of intersectingsubdomains:
Gi =M⋃
m=1
Gm
I ‘Walk-in-Subdomains’: Simulate exit point separately in every Gm
1. x0 = x , x1, . . . , xN – Markov chain, every xi+1 is an exit point fromthe corresponding subdomain for Brownian motion starting at xi
2. For spherical subdomains, B(xmi ,R
mi ), exit points are distributed in
accordance with the Poisson kernel:
14πRm
i
|xi − xmi |2 − (Rm
i )2
|xi − xi+1|3
![Page 99: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/99.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Fast Exit Point Calculations
Walk-in-Subdomains
‘Walk-on-Spheres’ and ‘Walk-in-Subdomains’I For general domains, an efficient way to simulate exit points is a
combination of1. Inside the molecule: ‘walk-in-subdomains’2. Outside the molecule ‘walk-on-spheres’
I The whole domain, Gi , is represented as a union of intersectingsubdomains:
Gi =M⋃
m=1
Gm
I ‘Walk-in-Subdomains’: Simulate exit point separately in every Gm
1. x0 = x , x1, . . . , xN – Markov chain, every xi+1 is an exit point fromthe corresponding subdomain for Brownian motion starting at xi
2. For spherical subdomains, B(xmi ,R
mi ), exit points are distributed in
accordance with the Poisson kernel:
14πRm
i
|xi − xmi |2 − (Rm
i )2
|xi − xi+1|3
![Page 100: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/100.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Fast Exit Point Calculations
Walk-in-Subdomains
‘Walk-on-Spheres’ and ‘Walk-in-Subdomains’I For general domains, an efficient way to simulate exit points is a
combination of1. Inside the molecule: ‘walk-in-subdomains’2. Outside the molecule ‘walk-on-spheres’
I The whole domain, Gi , is represented as a union of intersectingsubdomains:
Gi =M⋃
m=1
Gm
I ‘Walk-in-Subdomains’: Simulate exit point separately in every Gm
1. x0 = x , x1, . . . , xN – Markov chain, every xi+1 is an exit point fromthe corresponding subdomain for Brownian motion starting at xi
2. For spherical subdomains, B(xmi ,R
mi ), exit points are distributed in
accordance with the Poisson kernel:
14πRm
i
|xi − xmi |2 − (Rm
i )2
|xi − xi+1|3
![Page 101: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/101.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Fast Exit Point Calculations
Walk-in-Subdomains
‘Walk-on-Spheres’ and ‘Walk-in-Subdomains’I For general domains, an efficient way to simulate exit points is a
combination of1. Inside the molecule: ‘walk-in-subdomains’2. Outside the molecule ‘walk-on-spheres’
I The whole domain, Gi , is represented as a union of intersectingsubdomains:
Gi =M⋃
m=1
Gm
I ‘Walk-in-Subdomains’: Simulate exit point separately in every Gm
1. x0 = x , x1, . . . , xN – Markov chain, every xi+1 is an exit point fromthe corresponding subdomain for Brownian motion starting at xi
2. For spherical subdomains, B(xmi ,R
mi ), exit points are distributed in
accordance with the Poisson kernel:
14πRm
i
|xi − xmi |2 − (Rm
i )2
|xi − xi+1|3
![Page 102: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/102.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Fast Exit Point Calculations
Walk-in-Subdomains
‘Walk-on-Spheres’ and ‘Walk-in-Subdomains’I For general domains, an efficient way to simulate exit points is a
combination of1. Inside the molecule: ‘walk-in-subdomains’2. Outside the molecule ‘walk-on-spheres’
I The whole domain, Gi , is represented as a union of intersectingsubdomains:
Gi =M⋃
m=1
Gm
I ‘Walk-in-Subdomains’: Simulate exit point separately in every Gm
1. x0 = x , x1, . . . , xN – Markov chain, every xi+1 is an exit point fromthe corresponding subdomain for Brownian motion starting at xi
2. For spherical subdomains, B(xmi ,R
mi ), exit points are distributed in
accordance with the Poisson kernel:
14πRm
i
|xi − xmi |2 − (Rm
i )2
|xi − xi+1|3
![Page 103: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/103.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Fast Exit Point Calculations
Walk-in-Subdomains
‘Walk-on-Spheres’ and ‘Walk-in-Subdomains’I For general domains, an efficient way to simulate exit points is a
combination of1. Inside the molecule: ‘walk-in-subdomains’2. Outside the molecule ‘walk-on-spheres’
I The whole domain, Gi , is represented as a union of intersectingsubdomains:
Gi =M⋃
m=1
Gm
I ‘Walk-in-Subdomains’: Simulate exit point separately in every Gm
1. x0 = x , x1, . . . , xN – Markov chain, every xi+1 is an exit point fromthe corresponding subdomain for Brownian motion starting at xi
2. For spherical subdomains, B(xmi ,R
mi ), exit points are distributed in
accordance with the Poisson kernel:
14πRm
i
|xi − xmi |2 − (Rm
i )2
|xi − xi+1|3
![Page 104: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/104.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Fast Exit Point Calculations
Walk-in-Subdomains
‘Walk-on-Spheres’ and ‘Walk-in-Subdomains’
Figure: Walk in subdomains example
![Page 105: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/105.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Fast Exit Point Calculations
Monte Carlo Treatment of Boundary Conditions
Monte Carlo Treatment of Boundary Conditions
I Randomization of finite-difference approximation with step, h.u(y) = Eu(x) + O(h2)
I Exact treatment of boundary conditions (mean-value theorem)for boundary point, y , in the ball B(y ,a) with surface S(y ,a):
u(y) =εe
εe + εi
∫Se(y,a)
12πa2
κasinh(κa)
ue
+εi
εe + εi
∫Si (y,a)
12πa2
κasinh(κa)
ui (22)
− εe − εiεe + εi
∫Γ
⋂B(y,a)\y
cosϕyx
2π|y − x |2Qκ,au
+εi
εe + εi
∫Bi (y,a)
[−2κ2Φκ]ui
![Page 106: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/106.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Fast Exit Point Calculations
Monte Carlo Treatment of Boundary Conditions
Monte Carlo Treatment of Boundary Conditions
I Randomization of finite-difference approximation with step, h.u(y) = Eu(x) + O(h2)
I Exact treatment of boundary conditions (mean-value theorem)for boundary point, y , in the ball B(y ,a) with surface S(y ,a):
u(y) =εe
εe + εi
∫Se(y,a)
12πa2
κasinh(κa)
ue
+εi
εe + εi
∫Si (y,a)
12πa2
κasinh(κa)
ui (22)
− εe − εiεe + εi
∫Γ
⋂B(y,a)\y
cosϕyx
2π|y − x |2Qκ,au
+εi
εe + εi
∫Bi (y,a)
[−2κ2Φκ]ui
![Page 107: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/107.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Fast Exit Point Calculations
Monte Carlo Treatment of Boundary Conditions
Monte Carlo Treatment of Boundary Conditions
Randomized approximation to (22): u(y) = Eu(x) + O((a/2R)3):I With probability pe exit to solvent:
1. x is chosen isotropically on the surface of auxiliary sphere,S+(y , a), that lies above tangent plane
2. Walker survives with probabilityκa
sinh(κa)
I With probability pi = 1− pe:1. x is chosen isotropically in the solid angle below tangent plane;
with probability −2κ2Φκ & sampled in Bi (y , a) (reenter molecule)2. With the complementary probability x is sampled on the surface of
auxiliary sphere, S−(y , a), that lies below tangent plane3. x reenters molecule with conditional probability 1− a/2R and4. x exits to solvent with conditional probability a/2R
![Page 108: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/108.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Fast Exit Point Calculations
Monte Carlo Treatment of Boundary Conditions
Monte Carlo Treatment of Boundary Conditions
Randomized approximation to (22): u(y) = Eu(x) + O((a/2R)3):I With probability pe exit to solvent:
1. x is chosen isotropically on the surface of auxiliary sphere,S+(y , a), that lies above tangent plane
2. Walker survives with probabilityκa
sinh(κa)
I With probability pi = 1− pe:1. x is chosen isotropically in the solid angle below tangent plane;
with probability −2κ2Φκ & sampled in Bi (y , a) (reenter molecule)2. With the complementary probability x is sampled on the surface of
auxiliary sphere, S−(y , a), that lies below tangent plane3. x reenters molecule with conditional probability 1− a/2R and4. x exits to solvent with conditional probability a/2R
![Page 109: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/109.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Fast Exit Point Calculations
Monte Carlo Treatment of Boundary Conditions
Monte Carlo Treatment of Boundary Conditions
Randomized approximation to (22): u(y) = Eu(x) + O((a/2R)3):I With probability pe exit to solvent:
1. x is chosen isotropically on the surface of auxiliary sphere,S+(y , a), that lies above tangent plane
2. Walker survives with probabilityκa
sinh(κa)
I With probability pi = 1− pe:1. x is chosen isotropically in the solid angle below tangent plane;
with probability −2κ2Φκ & sampled in Bi (y , a) (reenter molecule)2. With the complementary probability x is sampled on the surface of
auxiliary sphere, S−(y , a), that lies below tangent plane3. x reenters molecule with conditional probability 1− a/2R and4. x exits to solvent with conditional probability a/2R
![Page 110: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/110.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Fast Exit Point Calculations
Monte Carlo Treatment of Boundary Conditions
Monte Carlo Treatment of Boundary Conditions
Randomized approximation to (22): u(y) = Eu(x) + O((a/2R)3):I With probability pe exit to solvent:
1. x is chosen isotropically on the surface of auxiliary sphere,S+(y , a), that lies above tangent plane
2. Walker survives with probabilityκa
sinh(κa)
I With probability pi = 1− pe:1. x is chosen isotropically in the solid angle below tangent plane;
with probability −2κ2Φκ & sampled in Bi (y , a) (reenter molecule)2. With the complementary probability x is sampled on the surface of
auxiliary sphere, S−(y , a), that lies below tangent plane3. x reenters molecule with conditional probability 1− a/2R and4. x exits to solvent with conditional probability a/2R
![Page 111: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/111.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Fast Exit Point Calculations
Monte Carlo Treatment of Boundary Conditions
Monte Carlo Treatment of Boundary Conditions
Randomized approximation to (22): u(y) = Eu(x) + O((a/2R)3):I With probability pe exit to solvent:
1. x is chosen isotropically on the surface of auxiliary sphere,S+(y , a), that lies above tangent plane
2. Walker survives with probabilityκa
sinh(κa)
I With probability pi = 1− pe:1. x is chosen isotropically in the solid angle below tangent plane;
with probability −2κ2Φκ & sampled in Bi (y , a) (reenter molecule)2. With the complementary probability x is sampled on the surface of
auxiliary sphere, S−(y , a), that lies below tangent plane3. x reenters molecule with conditional probability 1− a/2R and4. x exits to solvent with conditional probability a/2R
![Page 112: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/112.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Fast Exit Point Calculations
Monte Carlo Treatment of Boundary Conditions
Monte Carlo Treatment of Boundary Conditions
Randomized approximation to (22): u(y) = Eu(x) + O((a/2R)3):I With probability pe exit to solvent:
1. x is chosen isotropically on the surface of auxiliary sphere,S+(y , a), that lies above tangent plane
2. Walker survives with probabilityκa
sinh(κa)
I With probability pi = 1− pe:1. x is chosen isotropically in the solid angle below tangent plane;
with probability −2κ2Φκ & sampled in Bi (y , a) (reenter molecule)2. With the complementary probability x is sampled on the surface of
auxiliary sphere, S−(y , a), that lies below tangent plane3. x reenters molecule with conditional probability 1− a/2R and4. x exits to solvent with conditional probability a/2R
![Page 113: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/113.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Fast Exit Point Calculations
Monte Carlo Treatment of Boundary Conditions
Monte Carlo Treatment of Boundary Conditions
Randomized approximation to (22): u(y) = Eu(x) + O((a/2R)3):I With probability pe exit to solvent:
1. x is chosen isotropically on the surface of auxiliary sphere,S+(y , a), that lies above tangent plane
2. Walker survives with probabilityκa
sinh(κa)
I With probability pi = 1− pe:1. x is chosen isotropically in the solid angle below tangent plane;
with probability −2κ2Φκ & sampled in Bi (y , a) (reenter molecule)2. With the complementary probability x is sampled on the surface of
auxiliary sphere, S−(y , a), that lies below tangent plane3. x reenters molecule with conditional probability 1− a/2R and4. x exits to solvent with conditional probability a/2R
![Page 114: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/114.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Fast Exit Point Calculations
Monte Carlo Treatment of Boundary Conditions
Monte Carlo Treatment of Boundary Conditions
Randomized approximation to (22): u(y) = Eu(x) + O((a/2R)3):I With probability pe exit to solvent:
1. x is chosen isotropically on the surface of auxiliary sphere,S+(y , a), that lies above tangent plane
2. Walker survives with probabilityκa
sinh(κa)
I With probability pi = 1− pe:1. x is chosen isotropically in the solid angle below tangent plane;
with probability −2κ2Φκ & sampled in Bi (y , a) (reenter molecule)2. With the complementary probability x is sampled on the surface of
auxiliary sphere, S−(y , a), that lies below tangent plane3. x reenters molecule with conditional probability 1− a/2R and4. x exits to solvent with conditional probability a/2R
![Page 115: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/115.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Fast Exit Point Calculations
Monte Carlo Treatment of Boundary Conditions
Monte Carlo Treatment of Boundary Conditions
In the exterior, probability of terminating Markov chain dependslinearly on the initial distance to the boundary, d0. Therefore,⇒Mean number of returns to the boundary is O(d0)−1
I Finite-difference approximation of boundary conditions, ε = h2
Mean number of steps in the algorithm is O(h−1 log(h) f (κ)), f isa decreasing function (f (κ) = O(log(κ)) for small κ). Estimatesfor point values of the potential and free energy are O(h)-biased
I New treatment of boundary conditions provides O(a)2-biasedand more efficient Monte Carlo algorithm. Mean number ofsteps is O((a)−1 log(a) f (κ)), a = a/2R.
I More subtle approximation to (22) will provide even more efficientMonte Carlo estimates
![Page 116: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/116.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Fast Exit Point Calculations
Monte Carlo Treatment of Boundary Conditions
Monte Carlo Treatment of Boundary Conditions
In the exterior, probability of terminating Markov chain dependslinearly on the initial distance to the boundary, d0. Therefore,⇒Mean number of returns to the boundary is O(d0)−1
I Finite-difference approximation of boundary conditions, ε = h2
Mean number of steps in the algorithm is O(h−1 log(h) f (κ)), f isa decreasing function (f (κ) = O(log(κ)) for small κ). Estimatesfor point values of the potential and free energy are O(h)-biased
I New treatment of boundary conditions provides O(a)2-biasedand more efficient Monte Carlo algorithm. Mean number ofsteps is O((a)−1 log(a) f (κ)), a = a/2R.
I More subtle approximation to (22) will provide even more efficientMonte Carlo estimates
![Page 117: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/117.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Fast Exit Point Calculations
Monte Carlo Treatment of Boundary Conditions
Monte Carlo Treatment of Boundary Conditions
In the exterior, probability of terminating Markov chain dependslinearly on the initial distance to the boundary, d0. Therefore,⇒Mean number of returns to the boundary is O(d0)−1
I Finite-difference approximation of boundary conditions, ε = h2
Mean number of steps in the algorithm is O(h−1 log(h) f (κ)), f isa decreasing function (f (κ) = O(log(κ)) for small κ). Estimatesfor point values of the potential and free energy are O(h)-biased
I New treatment of boundary conditions provides O(a)2-biasedand more efficient Monte Carlo algorithm. Mean number ofsteps is O((a)−1 log(a) f (κ)), a = a/2R.
I More subtle approximation to (22) will provide even more efficientMonte Carlo estimates
![Page 118: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/118.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Monte Carlo Estimates
Monte Carlo Estimates
Monte Carlo Estimates
I The estimate for the reaction-field potential point value:ξ[Φrf ](x (m)) = −Φc(x∗1 )
+
Nins∑j=2
Fj (κ) (Φc(x insj )− Φc(x∗j,ins)) (23)
I Here x∗j,ins is a sequence of boundary points, after which therandom walker moves inside the domain, Gi , to x ins
jI The estimate for the reaction-field energy:
ξ[Wrf ] =12
M∑m=1
Qm ξ[Φrf ](x (m)) (24)
![Page 119: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/119.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Monte Carlo Estimates
Monte Carlo Estimates
Monte Carlo Estimates
I The estimate for the reaction-field potential point value:ξ[Φrf ](x (m)) = −Φc(x∗1 )
+
Nins∑j=2
Fj (κ) (Φc(x insj )− Φc(x∗j,ins)) (23)
I Here x∗j,ins is a sequence of boundary points, after which therandom walker moves inside the domain, Gi , to x ins
jI The estimate for the reaction-field energy:
ξ[Wrf ] =12
M∑m=1
Qm ξ[Φrf ](x (m)) (24)
![Page 120: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/120.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Monte Carlo Estimates
Monte Carlo Estimates
Monte Carlo Estimates
I The estimate for the reaction-field potential point value:ξ[Φrf ](x (m)) = −Φc(x∗1 )
+
Nins∑j=2
Fj (κ) (Φc(x insj )− Φc(x∗j,ins)) (23)
I Here x∗j,ins is a sequence of boundary points, after which therandom walker moves inside the domain, Gi , to x ins
jI The estimate for the reaction-field energy:
ξ[Wrf ] =12
M∑m=1
Qm ξ[Φrf ](x (m)) (24)
![Page 121: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/121.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Monte Carlo Estimates
Monte Carlo Estimates
A Picture: The Algorithm for a Single Spherical Atom
![Page 122: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/122.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Monte Carlo Estimates
Monte Carlo Estimates
The Algorithm in Pictures: Walk Inside
![Page 123: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/123.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Monte Carlo Estimates
Monte Carlo Estimates
The Algorithm in Pictures: Walk Inside
![Page 124: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/124.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Monte Carlo Estimates
Monte Carlo Estimates
The Algorithm in Pictures: Walk Outside
![Page 125: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/125.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Monte Carlo Estimates
Monte Carlo Estimates
The Algorithm in Pictures: Walk Outside
![Page 126: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/126.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Monte Carlo Estimates
Monte Carlo Estimates
The Algorithm in Pictures: Walk to∞ in One Step
Figure: κ = 0, p∞ = 1− REnclosed/dist
![Page 127: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/127.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Monte Carlo Estimates
Monte Carlo Estimates
Monte Carlo Algorithm’s Computational ComplexityCost of a single trajectory
I Number of steps is random walk is not dependent on M, thenumber of atoms
I The cost of finding the nearest sphere is M log2(M) due tooptimizations
0
0.5
1
1.5
2
0 1000 2000 3000 4000 5000
CP
UTi
me(
sec)
Number of Atoms
"Simulation""Theory-prediction"
Figure: The CPU time per atom per trajectory is plotted as function of number of atoms. For smallnumber of atoms the CPU time scales linearly and for large number of atoms it asymptoticallyscales logarithmically
![Page 128: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/128.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Monte Carlo Estimates
Monte Carlo Estimates
Monte Carlo Algorithm’s Computational ComplexityCost of a single trajectory
I Number of steps is random walk is not dependent on M, thenumber of atoms
I The cost of finding the nearest sphere is M log2(M) due tooptimizations
0
0.5
1
1.5
2
0 1000 2000 3000 4000 5000
CP
UTi
me(
sec)
Number of Atoms
"Simulation""Theory-prediction"
Figure: The CPU time per atom per trajectory is plotted as function of number of atoms. For smallnumber of atoms the CPU time scales linearly and for large number of atoms it asymptoticallyscales logarithmically
![Page 129: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/129.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Monte Carlo Estimates
Computational Geometry
Geometry: Problem Descriptions
There are many geometric problems that arise in this algorithm:
I Efficiently determining if a point is on the surface of the moleculeor inside of it (for interior walks)
I Efficiently determining the closest sphere to a given exterior point(for walks outside molecule)
I Efficiently determining if a query point is inside of the convex hullof the molecule
I Efficiently finding the largest possible sphere enclosing a querypoint for external walks
![Page 130: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/130.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Monte Carlo Estimates
Computational Geometry
Geometry: Problem Descriptions
There are many geometric problems that arise in this algorithm:
I Efficiently determining if a point is on the surface of the moleculeor inside of it (for interior walks)
I Efficiently determining the closest sphere to a given exterior point(for walks outside molecule)
I Efficiently determining if a query point is inside of the convex hullof the molecule
I Efficiently finding the largest possible sphere enclosing a querypoint for external walks
![Page 131: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/131.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Monte Carlo Estimates
Computational Geometry
Geometry: Problem Descriptions
There are many geometric problems that arise in this algorithm:
I Efficiently determining if a point is on the surface of the moleculeor inside of it (for interior walks)
I Efficiently determining the closest sphere to a given exterior point(for walks outside molecule)
I Efficiently determining if a query point is inside of the convex hullof the molecule
I Efficiently finding the largest possible sphere enclosing a querypoint for external walks
![Page 132: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/132.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Monte Carlo Estimates
Computational Geometry
Geometry: Problem Descriptions
There are many geometric problems that arise in this algorithm:
I Efficiently determining if a point is on the surface of the moleculeor inside of it (for interior walks)
I Efficiently determining the closest sphere to a given exterior point(for walks outside molecule)
I Efficiently determining if a query point is inside of the convex hullof the molecule
I Efficiently finding the largest possible sphere enclosing a querypoint for external walks
![Page 133: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/133.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Monte Carlo Estimates
Correlated and Uncorrelated Sampling
Correlated and Uncorrelated Sampling
I Correlated sampling in Monte Carlo is essential for two importantreasons
1. To obtain smooth curves with a minimum of sampling(function-wise vs. point-wise sampling)
2. To obtain accurate results from quantities defined as thedifferences of Monte Carlo estimates
I With this correlated sampling sampling you can get a “smoothcurve" with three orders of magnitude less sampling, note: youstill have O(N−1/2) errors, just in “curve space," not point by point
![Page 134: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/134.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Monte Carlo Estimates
Correlated and Uncorrelated Sampling
Correlated and Uncorrelated Sampling
I Correlated sampling in Monte Carlo is essential for two importantreasons
1. To obtain smooth curves with a minimum of sampling(function-wise vs. point-wise sampling)
2. To obtain accurate results from quantities defined as thedifferences of Monte Carlo estimates
I With this correlated sampling sampling you can get a “smoothcurve" with three orders of magnitude less sampling, note: youstill have O(N−1/2) errors, just in “curve space," not point by point
![Page 135: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/135.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Monte Carlo Estimates
Correlated and Uncorrelated Sampling
Correlated and Uncorrelated Sampling
I Correlated sampling in Monte Carlo is essential for two importantreasons
1. To obtain smooth curves with a minimum of sampling(function-wise vs. point-wise sampling)
2. To obtain accurate results from quantities defined as thedifferences of Monte Carlo estimates
I With this correlated sampling sampling you can get a “smoothcurve" with three orders of magnitude less sampling, note: youstill have O(N−1/2) errors, just in “curve space," not point by point
![Page 136: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/136.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Monte Carlo Estimates
Correlated and Uncorrelated Sampling
Correlated and Uncorrelated Sampling
I Correlated sampling in Monte Carlo is essential for two importantreasons
1. To obtain smooth curves with a minimum of sampling(function-wise vs. point-wise sampling)
2. To obtain accurate results from quantities defined as thedifferences of Monte Carlo estimates
I With this correlated sampling sampling you can get a “smoothcurve" with three orders of magnitude less sampling, note: youstill have O(N−1/2) errors, just in “curve space," not point by point
![Page 137: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/137.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Monte Carlo Estimates
Correlated and Uncorrelated Sampling
Correlated Sampling: Salt Concentration
-2970
-2965
-2960
-2955
-2950
-2945
-2940
-2935
-2930
-2925
-4 -3.5 -3 -2.5 -2 -1.5 -1 -0.5 0
∆G
elec
solv
(kca
l/mol
e)
log(Salt Concentration(M))
uncorr 500uncorr 1500
corr 500uncorr 4500
Figure: Electrostatic Solvation free Energy of 3icb calculated with three four conditions:uncorrelated sampling with 500 number of trajectories per concentration, uncorrelated samplingwith 1500 number of trajectories per concentration, uncorrelated sampling with 4500 number ofiterations, and correlated sampling with 500 number of trajectories
![Page 138: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/138.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Monte Carlo Estimates
Correlated and Uncorrelated Sampling
Dependence on Salt Concentration
I Values of scalar energies as a function of external saltconcentration are important
1. Smooth curves of internal energy vs. salt concentration (see above)2. Numerical estimate of the derivative as salt concentration vanishes
I For κ used in simulations, Fj (κ) = 1I For an arbitrary κ′ > κ:
Fj (κ′) is multiplied by the ratio
q(κ′,d)
q(κ,d)on every step of the WOS
in the exteriorI The results obtained with the estimates (23) and (24) for different
values of κ are highly correlated
![Page 139: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/139.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Monte Carlo Estimates
Correlated and Uncorrelated Sampling
Dependence on Salt Concentration
I Values of scalar energies as a function of external saltconcentration are important
1. Smooth curves of internal energy vs. salt concentration (see above)2. Numerical estimate of the derivative as salt concentration vanishes
I For κ used in simulations, Fj (κ) = 1I For an arbitrary κ′ > κ:
Fj (κ′) is multiplied by the ratio
q(κ′,d)
q(κ,d)on every step of the WOS
in the exteriorI The results obtained with the estimates (23) and (24) for different
values of κ are highly correlated
![Page 140: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/140.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Monte Carlo Estimates
Correlated and Uncorrelated Sampling
Dependence on Salt Concentration
I Values of scalar energies as a function of external saltconcentration are important
1. Smooth curves of internal energy vs. salt concentration (see above)2. Numerical estimate of the derivative as salt concentration vanishes
I For κ used in simulations, Fj (κ) = 1I For an arbitrary κ′ > κ:
Fj (κ′) is multiplied by the ratio
q(κ′,d)
q(κ,d)on every step of the WOS
in the exteriorI The results obtained with the estimates (23) and (24) for different
values of κ are highly correlated
![Page 141: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/141.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Monte Carlo Estimates
Correlated and Uncorrelated Sampling
Dependence on Salt Concentration
I Values of scalar energies as a function of external saltconcentration are important
1. Smooth curves of internal energy vs. salt concentration (see above)2. Numerical estimate of the derivative as salt concentration vanishes
I For κ used in simulations, Fj (κ) = 1I For an arbitrary κ′ > κ:
Fj (κ′) is multiplied by the ratio
q(κ′,d)
q(κ,d)on every step of the WOS
in the exteriorI The results obtained with the estimates (23) and (24) for different
values of κ are highly correlated
![Page 142: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/142.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Monte Carlo Estimates
Correlated and Uncorrelated Sampling
Dependence on Salt Concentration
I Values of scalar energies as a function of external saltconcentration are important
1. Smooth curves of internal energy vs. salt concentration (see above)2. Numerical estimate of the derivative as salt concentration vanishes
I For κ used in simulations, Fj (κ) = 1I For an arbitrary κ′ > κ:
Fj (κ′) is multiplied by the ratio
q(κ′,d)
q(κ,d)on every step of the WOS
in the exteriorI The results obtained with the estimates (23) and (24) for different
values of κ are highly correlated
![Page 143: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/143.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Monte Carlo Estimates
Correlated and Uncorrelated Sampling
Dependence on Salt Concentration
I Values of scalar energies as a function of external saltconcentration are important
1. Smooth curves of internal energy vs. salt concentration (see above)2. Numerical estimate of the derivative as salt concentration vanishes
I For κ used in simulations, Fj (κ) = 1I For an arbitrary κ′ > κ:
Fj (κ′) is multiplied by the ratio
q(κ′,d)
q(κ,d)on every step of the WOS
in the exteriorI The results obtained with the estimates (23) and (24) for different
values of κ are highly correlated
![Page 144: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/144.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Monte Carlo Estimates
Correlated and Uncorrelated Sampling
Correlated Sampling: Binding Calculations
I Binding computation requires three energy computationsE(A + B)− E(A)− E(B)
I Monte Carlo requires “help" when differencingI We use the reproducibility in SPRNG to do this effectively
1. Unbound: when exiting the molecule the seed is stored usingSPRNG tools
2. Bound: walks resume at the exit points with the same randomnumber streams and reusing
3. At this exit point, only the exit point information is requiredI The leads to correlation between unbound and bound energy
computations that decreases as the walk length increases (κ2
decreases)
![Page 145: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/145.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Monte Carlo Estimates
Correlated and Uncorrelated Sampling
Correlated Sampling: Binding Calculations
I Binding computation requires three energy computationsE(A + B)− E(A)− E(B)
I Monte Carlo requires “help" when differencingI We use the reproducibility in SPRNG to do this effectively
1. Unbound: when exiting the molecule the seed is stored usingSPRNG tools
2. Bound: walks resume at the exit points with the same randomnumber streams and reusing
3. At this exit point, only the exit point information is requiredI The leads to correlation between unbound and bound energy
computations that decreases as the walk length increases (κ2
decreases)
![Page 146: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/146.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Monte Carlo Estimates
Correlated and Uncorrelated Sampling
Correlated Sampling: Binding Calculations
I Binding computation requires three energy computationsE(A + B)− E(A)− E(B)
I Monte Carlo requires “help" when differencingI We use the reproducibility in SPRNG to do this effectively
1. Unbound: when exiting the molecule the seed is stored usingSPRNG tools
2. Bound: walks resume at the exit points with the same randomnumber streams and reusing
3. At this exit point, only the exit point information is requiredI The leads to correlation between unbound and bound energy
computations that decreases as the walk length increases (κ2
decreases)
![Page 147: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/147.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Monte Carlo Estimates
Correlated and Uncorrelated Sampling
Correlated Sampling: Binding Calculations
I Binding computation requires three energy computationsE(A + B)− E(A)− E(B)
I Monte Carlo requires “help" when differencingI We use the reproducibility in SPRNG to do this effectively
1. Unbound: when exiting the molecule the seed is stored usingSPRNG tools
2. Bound: walks resume at the exit points with the same randomnumber streams and reusing
3. At this exit point, only the exit point information is requiredI The leads to correlation between unbound and bound energy
computations that decreases as the walk length increases (κ2
decreases)
![Page 148: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/148.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Monte Carlo Estimates
Correlated and Uncorrelated Sampling
Correlated Sampling: Binding Calculations
I Binding computation requires three energy computationsE(A + B)− E(A)− E(B)
I Monte Carlo requires “help" when differencingI We use the reproducibility in SPRNG to do this effectively
1. Unbound: when exiting the molecule the seed is stored usingSPRNG tools
2. Bound: walks resume at the exit points with the same randomnumber streams and reusing
3. At this exit point, only the exit point information is requiredI The leads to correlation between unbound and bound energy
computations that decreases as the walk length increases (κ2
decreases)
![Page 149: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/149.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Monte Carlo Estimates
Correlated and Uncorrelated Sampling
Correlated Sampling: Binding Calculations
I Binding computation requires three energy computationsE(A + B)− E(A)− E(B)
I Monte Carlo requires “help" when differencingI We use the reproducibility in SPRNG to do this effectively
1. Unbound: when exiting the molecule the seed is stored usingSPRNG tools
2. Bound: walks resume at the exit points with the same randomnumber streams and reusing
3. At this exit point, only the exit point information is requiredI The leads to correlation between unbound and bound energy
computations that decreases as the walk length increases (κ2
decreases)
![Page 150: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/150.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Monte Carlo Estimates
Correlated and Uncorrelated Sampling
Correlated Sampling: Binding Calculations
I Binding computation requires three energy computationsE(A + B)− E(A)− E(B)
I Monte Carlo requires “help" when differencingI We use the reproducibility in SPRNG to do this effectively
1. Unbound: when exiting the molecule the seed is stored usingSPRNG tools
2. Bound: walks resume at the exit points with the same randomnumber streams and reusing
3. At this exit point, only the exit point information is requiredI The leads to correlation between unbound and bound energy
computations that decreases as the walk length increases (κ2
decreases)
![Page 151: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/151.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Computational Results
Accuracy: Monte Carlo vs. Deterministic
![Page 152: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/152.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Computational Results
Sampling Error and Bias
I In Monte Carlo there are biases (errors) and sampling error1. Sampling error is based on standard error O(N−1/2)2. Difference between expected value and PDE solution is bias
I Capture thickness (ε): bias is O(ε)I Auxiliary sphere radius (a): bias is O(a3)I Effective Van der Waals sphere radius, RI Overall bias:
( a2R
)3+
(ε
2R
)3. Var [
∑i qi Φ(xi )] =
∑i q2
i Var [Φ(xi )]4. Given a desired variance, divide it evenly over this sum5. Running time ∝ | ln(ε)|
a6. Can reduce running time by 2 orders of magnitude by bias/variance
balancing and using larger ε, a and ANN7. Large ANN means errors in drawing the largest sphere outside the
molecule for WOS
![Page 153: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/153.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Computational Results
Sampling Error and Bias
I In Monte Carlo there are biases (errors) and sampling error1. Sampling error is based on standard error O(N−1/2)2. Difference between expected value and PDE solution is bias
I Capture thickness (ε): bias is O(ε)I Auxiliary sphere radius (a): bias is O(a3)I Effective Van der Waals sphere radius, RI Overall bias:
( a2R
)3+
(ε
2R
)3. Var [
∑i qi Φ(xi )] =
∑i q2
i Var [Φ(xi )]4. Given a desired variance, divide it evenly over this sum5. Running time ∝ | ln(ε)|
a6. Can reduce running time by 2 orders of magnitude by bias/variance
balancing and using larger ε, a and ANN7. Large ANN means errors in drawing the largest sphere outside the
molecule for WOS
![Page 154: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/154.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Computational Results
Sampling Error and Bias
I In Monte Carlo there are biases (errors) and sampling error1. Sampling error is based on standard error O(N−1/2)2. Difference between expected value and PDE solution is bias
I Capture thickness (ε): bias is O(ε)I Auxiliary sphere radius (a): bias is O(a3)I Effective Van der Waals sphere radius, RI Overall bias:
( a2R
)3+
(ε
2R
)3. Var [
∑i qi Φ(xi )] =
∑i q2
i Var [Φ(xi )]4. Given a desired variance, divide it evenly over this sum5. Running time ∝ | ln(ε)|
a6. Can reduce running time by 2 orders of magnitude by bias/variance
balancing and using larger ε, a and ANN7. Large ANN means errors in drawing the largest sphere outside the
molecule for WOS
![Page 155: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/155.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Computational Results
Sampling Error and Bias
I In Monte Carlo there are biases (errors) and sampling error1. Sampling error is based on standard error O(N−1/2)2. Difference between expected value and PDE solution is bias
I Capture thickness (ε): bias is O(ε)I Auxiliary sphere radius (a): bias is O(a3)I Effective Van der Waals sphere radius, RI Overall bias:
( a2R
)3+
(ε
2R
)3. Var [
∑i qi Φ(xi )] =
∑i q2
i Var [Φ(xi )]4. Given a desired variance, divide it evenly over this sum5. Running time ∝ | ln(ε)|
a6. Can reduce running time by 2 orders of magnitude by bias/variance
balancing and using larger ε, a and ANN7. Large ANN means errors in drawing the largest sphere outside the
molecule for WOS
![Page 156: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/156.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Computational Results
Sampling Error and Bias
I In Monte Carlo there are biases (errors) and sampling error1. Sampling error is based on standard error O(N−1/2)2. Difference between expected value and PDE solution is bias
I Capture thickness (ε): bias is O(ε)I Auxiliary sphere radius (a): bias is O(a3)I Effective Van der Waals sphere radius, RI Overall bias:
( a2R
)3+
(ε
2R
)3. Var [
∑i qi Φ(xi )] =
∑i q2
i Var [Φ(xi )]4. Given a desired variance, divide it evenly over this sum5. Running time ∝ | ln(ε)|
a6. Can reduce running time by 2 orders of magnitude by bias/variance
balancing and using larger ε, a and ANN7. Large ANN means errors in drawing the largest sphere outside the
molecule for WOS
![Page 157: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/157.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Computational Results
Sampling Error and Bias
I In Monte Carlo there are biases (errors) and sampling error1. Sampling error is based on standard error O(N−1/2)2. Difference between expected value and PDE solution is bias
I Capture thickness (ε): bias is O(ε)I Auxiliary sphere radius (a): bias is O(a3)I Effective Van der Waals sphere radius, RI Overall bias:
( a2R
)3+
(ε
2R
)3. Var [
∑i qi Φ(xi )] =
∑i q2
i Var [Φ(xi )]4. Given a desired variance, divide it evenly over this sum5. Running time ∝ | ln(ε)|
a6. Can reduce running time by 2 orders of magnitude by bias/variance
balancing and using larger ε, a and ANN7. Large ANN means errors in drawing the largest sphere outside the
molecule for WOS
![Page 158: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/158.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Computational Results
Sampling Error and Bias
I In Monte Carlo there are biases (errors) and sampling error1. Sampling error is based on standard error O(N−1/2)2. Difference between expected value and PDE solution is bias
I Capture thickness (ε): bias is O(ε)I Auxiliary sphere radius (a): bias is O(a3)I Effective Van der Waals sphere radius, RI Overall bias:
( a2R
)3+
(ε
2R
)3. Var [
∑i qi Φ(xi )] =
∑i q2
i Var [Φ(xi )]4. Given a desired variance, divide it evenly over this sum5. Running time ∝ | ln(ε)|
a6. Can reduce running time by 2 orders of magnitude by bias/variance
balancing and using larger ε, a and ANN7. Large ANN means errors in drawing the largest sphere outside the
molecule for WOS
![Page 159: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/159.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Computational Results
Sampling Error and Bias
I In Monte Carlo there are biases (errors) and sampling error1. Sampling error is based on standard error O(N−1/2)2. Difference between expected value and PDE solution is bias
I Capture thickness (ε): bias is O(ε)I Auxiliary sphere radius (a): bias is O(a3)I Effective Van der Waals sphere radius, RI Overall bias:
( a2R
)3+
(ε
2R
)3. Var [
∑i qi Φ(xi )] =
∑i q2
i Var [Φ(xi )]4. Given a desired variance, divide it evenly over this sum5. Running time ∝ | ln(ε)|
a6. Can reduce running time by 2 orders of magnitude by bias/variance
balancing and using larger ε, a and ANN7. Large ANN means errors in drawing the largest sphere outside the
molecule for WOS
![Page 160: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/160.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Computational Results
Sampling Error and Bias
I In Monte Carlo there are biases (errors) and sampling error1. Sampling error is based on standard error O(N−1/2)2. Difference between expected value and PDE solution is bias
I Capture thickness (ε): bias is O(ε)I Auxiliary sphere radius (a): bias is O(a3)I Effective Van der Waals sphere radius, RI Overall bias:
( a2R
)3+
(ε
2R
)3. Var [
∑i qi Φ(xi )] =
∑i q2
i Var [Φ(xi )]4. Given a desired variance, divide it evenly over this sum5. Running time ∝ | ln(ε)|
a6. Can reduce running time by 2 orders of magnitude by bias/variance
balancing and using larger ε, a and ANN7. Large ANN means errors in drawing the largest sphere outside the
molecule for WOS
![Page 161: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/161.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Computational Results
Sampling Error and Bias
I In Monte Carlo there are biases (errors) and sampling error1. Sampling error is based on standard error O(N−1/2)2. Difference between expected value and PDE solution is bias
I Capture thickness (ε): bias is O(ε)I Auxiliary sphere radius (a): bias is O(a3)I Effective Van der Waals sphere radius, RI Overall bias:
( a2R
)3+
(ε
2R
)3. Var [
∑i qi Φ(xi )] =
∑i q2
i Var [Φ(xi )]4. Given a desired variance, divide it evenly over this sum5. Running time ∝ | ln(ε)|
a6. Can reduce running time by 2 orders of magnitude by bias/variance
balancing and using larger ε, a and ANN7. Large ANN means errors in drawing the largest sphere outside the
molecule for WOS
![Page 162: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/162.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Computational Results
Sampling Error and Bias
I In Monte Carlo there are biases (errors) and sampling error1. Sampling error is based on standard error O(N−1/2)2. Difference between expected value and PDE solution is bias
I Capture thickness (ε): bias is O(ε)I Auxiliary sphere radius (a): bias is O(a3)I Effective Van der Waals sphere radius, RI Overall bias:
( a2R
)3+
(ε
2R
)3. Var [
∑i qi Φ(xi )] =
∑i q2
i Var [Φ(xi )]4. Given a desired variance, divide it evenly over this sum5. Running time ∝ | ln(ε)|
a6. Can reduce running time by 2 orders of magnitude by bias/variance
balancing and using larger ε, a and ANN7. Large ANN means errors in drawing the largest sphere outside the
molecule for WOS
![Page 163: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/163.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Computational Results
Sampling Error and Bias
I In Monte Carlo there are biases (errors) and sampling error1. Sampling error is based on standard error O(N−1/2)2. Difference between expected value and PDE solution is bias
I Capture thickness (ε): bias is O(ε)I Auxiliary sphere radius (a): bias is O(a3)I Effective Van der Waals sphere radius, RI Overall bias:
( a2R
)3+
(ε
2R
)3. Var [
∑i qi Φ(xi )] =
∑i q2
i Var [Φ(xi )]4. Given a desired variance, divide it evenly over this sum5. Running time ∝ | ln(ε)|
a6. Can reduce running time by 2 orders of magnitude by bias/variance
balancing and using larger ε, a and ANN7. Large ANN means errors in drawing the largest sphere outside the
molecule for WOS
![Page 164: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/164.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Computational Results
Timing: Better Than Expected
Figure: O(M log M)?
![Page 165: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/165.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Conclusions and Future Work
ConclusionsI Over the years we have developed many MC tools for PDEs and
more recently:I We have developed a novel stochastic linear PBE solver that can
provide highly accurate salt-dependent electrostatic properties ofbiomolecules in a single PBE calculation
I Advantages of the stochastic linear PBE solver over the moremature deterministic methods include: the subtle geometricfeatures of the biomolecule can be treated with higher precision,the continuity and outer boundary conditions are accounted forexactly, a singularity free scheme is employed andstraightforward implementation on parallel computer platform ispossible
I Codes provide higher accuracy (on demand) and do not sufferlosses in accuracy near the boundary
I Only way to handle large (M >> 10000) molecules
![Page 166: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/166.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Conclusions and Future Work
ConclusionsI Over the years we have developed many MC tools for PDEs and
more recently:I We have developed a novel stochastic linear PBE solver that can
provide highly accurate salt-dependent electrostatic properties ofbiomolecules in a single PBE calculation
I Advantages of the stochastic linear PBE solver over the moremature deterministic methods include: the subtle geometricfeatures of the biomolecule can be treated with higher precision,the continuity and outer boundary conditions are accounted forexactly, a singularity free scheme is employed andstraightforward implementation on parallel computer platform ispossible
I Codes provide higher accuracy (on demand) and do not sufferlosses in accuracy near the boundary
I Only way to handle large (M >> 10000) molecules
![Page 167: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/167.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Conclusions and Future Work
ConclusionsI Over the years we have developed many MC tools for PDEs and
more recently:I We have developed a novel stochastic linear PBE solver that can
provide highly accurate salt-dependent electrostatic properties ofbiomolecules in a single PBE calculation
I Advantages of the stochastic linear PBE solver over the moremature deterministic methods include: the subtle geometricfeatures of the biomolecule can be treated with higher precision,the continuity and outer boundary conditions are accounted forexactly, a singularity free scheme is employed andstraightforward implementation on parallel computer platform ispossible
I Codes provide higher accuracy (on demand) and do not sufferlosses in accuracy near the boundary
I Only way to handle large (M >> 10000) molecules
![Page 168: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/168.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Conclusions and Future Work
ConclusionsI Over the years we have developed many MC tools for PDEs and
more recently:I We have developed a novel stochastic linear PBE solver that can
provide highly accurate salt-dependent electrostatic properties ofbiomolecules in a single PBE calculation
I Advantages of the stochastic linear PBE solver over the moremature deterministic methods include: the subtle geometricfeatures of the biomolecule can be treated with higher precision,the continuity and outer boundary conditions are accounted forexactly, a singularity free scheme is employed andstraightforward implementation on parallel computer platform ispossible
I Codes provide higher accuracy (on demand) and do not sufferlosses in accuracy near the boundary
I Only way to handle large (M >> 10000) molecules
![Page 169: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/169.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Conclusions and Future Work
ConclusionsI Over the years we have developed many MC tools for PDEs and
more recently:I We have developed a novel stochastic linear PBE solver that can
provide highly accurate salt-dependent electrostatic properties ofbiomolecules in a single PBE calculation
I Advantages of the stochastic linear PBE solver over the moremature deterministic methods include: the subtle geometricfeatures of the biomolecule can be treated with higher precision,the continuity and outer boundary conditions are accounted forexactly, a singularity free scheme is employed andstraightforward implementation on parallel computer platform ispossible
I Codes provide higher accuracy (on demand) and do not sufferlosses in accuracy near the boundary
I Only way to handle large (M >> 10000) molecules
![Page 170: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/170.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Conclusions and Future Work
Future WorkI Binding computations: using correlated sampling by directly
reprocessing walksI Simple code interface for distribution with
1. Desired accuracy as input that allows a precalculation of thenumber of needed trajectories
2. Importance sampling for optimal estimation of scalar energy values3. Built-in CONDOR support for distribution of concurrent tasks4. Multicore distributed computing support for the code:
OpenMP/OpenMPI5. Precompiled code module distribution to protect IP6. Webpage to describe the method and the mathematical
background and applicationI Exploit the implicit inverse computation this methods provides
1. Can do computation without knowing charges until the end (aninverse)
2. Simple to examine many charge distributions in a perfectlycorrelated setting
![Page 171: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/171.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Conclusions and Future Work
Future WorkI Binding computations: using correlated sampling by directly
reprocessing walksI Simple code interface for distribution with
1. Desired accuracy as input that allows a precalculation of thenumber of needed trajectories
2. Importance sampling for optimal estimation of scalar energy values3. Built-in CONDOR support for distribution of concurrent tasks4. Multicore distributed computing support for the code:
OpenMP/OpenMPI5. Precompiled code module distribution to protect IP6. Webpage to describe the method and the mathematical
background and applicationI Exploit the implicit inverse computation this methods provides
1. Can do computation without knowing charges until the end (aninverse)
2. Simple to examine many charge distributions in a perfectlycorrelated setting
![Page 172: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/172.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Conclusions and Future Work
Future WorkI Binding computations: using correlated sampling by directly
reprocessing walksI Simple code interface for distribution with
1. Desired accuracy as input that allows a precalculation of thenumber of needed trajectories
2. Importance sampling for optimal estimation of scalar energy values3. Built-in CONDOR support for distribution of concurrent tasks4. Multicore distributed computing support for the code:
OpenMP/OpenMPI5. Precompiled code module distribution to protect IP6. Webpage to describe the method and the mathematical
background and applicationI Exploit the implicit inverse computation this methods provides
1. Can do computation without knowing charges until the end (aninverse)
2. Simple to examine many charge distributions in a perfectlycorrelated setting
![Page 173: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/173.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Conclusions and Future Work
Future WorkI Binding computations: using correlated sampling by directly
reprocessing walksI Simple code interface for distribution with
1. Desired accuracy as input that allows a precalculation of thenumber of needed trajectories
2. Importance sampling for optimal estimation of scalar energy values3. Built-in CONDOR support for distribution of concurrent tasks4. Multicore distributed computing support for the code:
OpenMP/OpenMPI5. Precompiled code module distribution to protect IP6. Webpage to describe the method and the mathematical
background and applicationI Exploit the implicit inverse computation this methods provides
1. Can do computation without knowing charges until the end (aninverse)
2. Simple to examine many charge distributions in a perfectlycorrelated setting
![Page 174: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/174.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Conclusions and Future Work
Future WorkI Binding computations: using correlated sampling by directly
reprocessing walksI Simple code interface for distribution with
1. Desired accuracy as input that allows a precalculation of thenumber of needed trajectories
2. Importance sampling for optimal estimation of scalar energy values3. Built-in CONDOR support for distribution of concurrent tasks4. Multicore distributed computing support for the code:
OpenMP/OpenMPI5. Precompiled code module distribution to protect IP6. Webpage to describe the method and the mathematical
background and applicationI Exploit the implicit inverse computation this methods provides
1. Can do computation without knowing charges until the end (aninverse)
2. Simple to examine many charge distributions in a perfectlycorrelated setting
![Page 175: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/175.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Conclusions and Future Work
Future WorkI Binding computations: using correlated sampling by directly
reprocessing walksI Simple code interface for distribution with
1. Desired accuracy as input that allows a precalculation of thenumber of needed trajectories
2. Importance sampling for optimal estimation of scalar energy values3. Built-in CONDOR support for distribution of concurrent tasks4. Multicore distributed computing support for the code:
OpenMP/OpenMPI5. Precompiled code module distribution to protect IP6. Webpage to describe the method and the mathematical
background and applicationI Exploit the implicit inverse computation this methods provides
1. Can do computation without knowing charges until the end (aninverse)
2. Simple to examine many charge distributions in a perfectlycorrelated setting
![Page 176: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/176.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Conclusions and Future Work
Future WorkI Binding computations: using correlated sampling by directly
reprocessing walksI Simple code interface for distribution with
1. Desired accuracy as input that allows a precalculation of thenumber of needed trajectories
2. Importance sampling for optimal estimation of scalar energy values3. Built-in CONDOR support for distribution of concurrent tasks4. Multicore distributed computing support for the code:
OpenMP/OpenMPI5. Precompiled code module distribution to protect IP6. Webpage to describe the method and the mathematical
background and applicationI Exploit the implicit inverse computation this methods provides
1. Can do computation without knowing charges until the end (aninverse)
2. Simple to examine many charge distributions in a perfectlycorrelated setting
![Page 177: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/177.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Conclusions and Future Work
Future WorkI Binding computations: using correlated sampling by directly
reprocessing walksI Simple code interface for distribution with
1. Desired accuracy as input that allows a precalculation of thenumber of needed trajectories
2. Importance sampling for optimal estimation of scalar energy values3. Built-in CONDOR support for distribution of concurrent tasks4. Multicore distributed computing support for the code:
OpenMP/OpenMPI5. Precompiled code module distribution to protect IP6. Webpage to describe the method and the mathematical
background and applicationI Exploit the implicit inverse computation this methods provides
1. Can do computation without knowing charges until the end (aninverse)
2. Simple to examine many charge distributions in a perfectlycorrelated setting
![Page 178: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/178.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Conclusions and Future Work
Future WorkI Binding computations: using correlated sampling by directly
reprocessing walksI Simple code interface for distribution with
1. Desired accuracy as input that allows a precalculation of thenumber of needed trajectories
2. Importance sampling for optimal estimation of scalar energy values3. Built-in CONDOR support for distribution of concurrent tasks4. Multicore distributed computing support for the code:
OpenMP/OpenMPI5. Precompiled code module distribution to protect IP6. Webpage to describe the method and the mathematical
background and applicationI Exploit the implicit inverse computation this methods provides
1. Can do computation without knowing charges until the end (aninverse)
2. Simple to examine many charge distributions in a perfectlycorrelated setting
![Page 179: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/179.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Conclusions and Future Work
Future WorkI Binding computations: using correlated sampling by directly
reprocessing walksI Simple code interface for distribution with
1. Desired accuracy as input that allows a precalculation of thenumber of needed trajectories
2. Importance sampling for optimal estimation of scalar energy values3. Built-in CONDOR support for distribution of concurrent tasks4. Multicore distributed computing support for the code:
OpenMP/OpenMPI5. Precompiled code module distribution to protect IP6. Webpage to describe the method and the mathematical
background and applicationI Exploit the implicit inverse computation this methods provides
1. Can do computation without knowing charges until the end (aninverse)
2. Simple to examine many charge distributions in a perfectlycorrelated setting
![Page 180: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/180.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Conclusions and Future Work
Future WorkI Binding computations: using correlated sampling by directly
reprocessing walksI Simple code interface for distribution with
1. Desired accuracy as input that allows a precalculation of thenumber of needed trajectories
2. Importance sampling for optimal estimation of scalar energy values3. Built-in CONDOR support for distribution of concurrent tasks4. Multicore distributed computing support for the code:
OpenMP/OpenMPI5. Precompiled code module distribution to protect IP6. Webpage to describe the method and the mathematical
background and applicationI Exploit the implicit inverse computation this methods provides
1. Can do computation without knowing charges until the end (aninverse)
2. Simple to examine many charge distributions in a perfectlycorrelated setting
![Page 181: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/181.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Conclusions and Future Work
Future Work
I Further algorithmic development1. Computation of gradients using existing Markov chains2. Global computation of field variables and their visualization3. Nonlinear BVPs perhaps via branching processes4. Using “Walk-on-the-Boundary" (WOB) techniques
I Geometric Issues1. Computation of the three region model problem2. More complicated surfaces (solvent-excluded and ion-excluded)3. Accuracy issues related to the Van der Waals surface
I Optimize the performance1. Error/bias/variance balancing2. Importance sampling and the outer walks3. WOB to eliminate walks outside4. QMC methods
![Page 182: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/182.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Conclusions and Future Work
Future Work
I Further algorithmic development1. Computation of gradients using existing Markov chains2. Global computation of field variables and their visualization3. Nonlinear BVPs perhaps via branching processes4. Using “Walk-on-the-Boundary" (WOB) techniques
I Geometric Issues1. Computation of the three region model problem2. More complicated surfaces (solvent-excluded and ion-excluded)3. Accuracy issues related to the Van der Waals surface
I Optimize the performance1. Error/bias/variance balancing2. Importance sampling and the outer walks3. WOB to eliminate walks outside4. QMC methods
![Page 183: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/183.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Conclusions and Future Work
Future Work
I Further algorithmic development1. Computation of gradients using existing Markov chains2. Global computation of field variables and their visualization3. Nonlinear BVPs perhaps via branching processes4. Using “Walk-on-the-Boundary" (WOB) techniques
I Geometric Issues1. Computation of the three region model problem2. More complicated surfaces (solvent-excluded and ion-excluded)3. Accuracy issues related to the Van der Waals surface
I Optimize the performance1. Error/bias/variance balancing2. Importance sampling and the outer walks3. WOB to eliminate walks outside4. QMC methods
![Page 184: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/184.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Conclusions and Future Work
Future Work
I Further algorithmic development1. Computation of gradients using existing Markov chains2. Global computation of field variables and their visualization3. Nonlinear BVPs perhaps via branching processes4. Using “Walk-on-the-Boundary" (WOB) techniques
I Geometric Issues1. Computation of the three region model problem2. More complicated surfaces (solvent-excluded and ion-excluded)3. Accuracy issues related to the Van der Waals surface
I Optimize the performance1. Error/bias/variance balancing2. Importance sampling and the outer walks3. WOB to eliminate walks outside4. QMC methods
![Page 185: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/185.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Conclusions and Future Work
Future Work
I Further algorithmic development1. Computation of gradients using existing Markov chains2. Global computation of field variables and their visualization3. Nonlinear BVPs perhaps via branching processes4. Using “Walk-on-the-Boundary" (WOB) techniques
I Geometric Issues1. Computation of the three region model problem2. More complicated surfaces (solvent-excluded and ion-excluded)3. Accuracy issues related to the Van der Waals surface
I Optimize the performance1. Error/bias/variance balancing2. Importance sampling and the outer walks3. WOB to eliminate walks outside4. QMC methods
![Page 186: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/186.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Conclusions and Future Work
Future Work
I Further algorithmic development1. Computation of gradients using existing Markov chains2. Global computation of field variables and their visualization3. Nonlinear BVPs perhaps via branching processes4. Using “Walk-on-the-Boundary" (WOB) techniques
I Geometric Issues1. Computation of the three region model problem2. More complicated surfaces (solvent-excluded and ion-excluded)3. Accuracy issues related to the Van der Waals surface
I Optimize the performance1. Error/bias/variance balancing2. Importance sampling and the outer walks3. WOB to eliminate walks outside4. QMC methods
![Page 187: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/187.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Conclusions and Future Work
Future Work
I Further algorithmic development1. Computation of gradients using existing Markov chains2. Global computation of field variables and their visualization3. Nonlinear BVPs perhaps via branching processes4. Using “Walk-on-the-Boundary" (WOB) techniques
I Geometric Issues1. Computation of the three region model problem2. More complicated surfaces (solvent-excluded and ion-excluded)3. Accuracy issues related to the Van der Waals surface
I Optimize the performance1. Error/bias/variance balancing2. Importance sampling and the outer walks3. WOB to eliminate walks outside4. QMC methods
![Page 188: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/188.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Conclusions and Future Work
Future Work
I Further algorithmic development1. Computation of gradients using existing Markov chains2. Global computation of field variables and their visualization3. Nonlinear BVPs perhaps via branching processes4. Using “Walk-on-the-Boundary" (WOB) techniques
I Geometric Issues1. Computation of the three region model problem2. More complicated surfaces (solvent-excluded and ion-excluded)3. Accuracy issues related to the Van der Waals surface
I Optimize the performance1. Error/bias/variance balancing2. Importance sampling and the outer walks3. WOB to eliminate walks outside4. QMC methods
![Page 189: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/189.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Conclusions and Future Work
Future Work
I Further algorithmic development1. Computation of gradients using existing Markov chains2. Global computation of field variables and their visualization3. Nonlinear BVPs perhaps via branching processes4. Using “Walk-on-the-Boundary" (WOB) techniques
I Geometric Issues1. Computation of the three region model problem2. More complicated surfaces (solvent-excluded and ion-excluded)3. Accuracy issues related to the Van der Waals surface
I Optimize the performance1. Error/bias/variance balancing2. Importance sampling and the outer walks3. WOB to eliminate walks outside4. QMC methods
![Page 190: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/190.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Conclusions and Future Work
Future Work
I Further algorithmic development1. Computation of gradients using existing Markov chains2. Global computation of field variables and their visualization3. Nonlinear BVPs perhaps via branching processes4. Using “Walk-on-the-Boundary" (WOB) techniques
I Geometric Issues1. Computation of the three region model problem2. More complicated surfaces (solvent-excluded and ion-excluded)3. Accuracy issues related to the Van der Waals surface
I Optimize the performance1. Error/bias/variance balancing2. Importance sampling and the outer walks3. WOB to eliminate walks outside4. QMC methods
![Page 191: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/191.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Conclusions and Future Work
Future Work
I Further algorithmic development1. Computation of gradients using existing Markov chains2. Global computation of field variables and their visualization3. Nonlinear BVPs perhaps via branching processes4. Using “Walk-on-the-Boundary" (WOB) techniques
I Geometric Issues1. Computation of the three region model problem2. More complicated surfaces (solvent-excluded and ion-excluded)3. Accuracy issues related to the Van der Waals surface
I Optimize the performance1. Error/bias/variance balancing2. Importance sampling and the outer walks3. WOB to eliminate walks outside4. QMC methods
![Page 192: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/192.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Conclusions and Future Work
Future Work
I Further algorithmic development1. Computation of gradients using existing Markov chains2. Global computation of field variables and their visualization3. Nonlinear BVPs perhaps via branching processes4. Using “Walk-on-the-Boundary" (WOB) techniques
I Geometric Issues1. Computation of the three region model problem2. More complicated surfaces (solvent-excluded and ion-excluded)3. Accuracy issues related to the Van der Waals surface
I Optimize the performance1. Error/bias/variance balancing2. Importance sampling and the outer walks3. WOB to eliminate walks outside4. QMC methods
![Page 193: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/193.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Conclusions and Future Work
Future Work
I Further algorithmic development1. Computation of gradients using existing Markov chains2. Global computation of field variables and their visualization3. Nonlinear BVPs perhaps via branching processes4. Using “Walk-on-the-Boundary" (WOB) techniques
I Geometric Issues1. Computation of the three region model problem2. More complicated surfaces (solvent-excluded and ion-excluded)3. Accuracy issues related to the Van der Waals surface
I Optimize the performance1. Error/bias/variance balancing2. Importance sampling and the outer walks3. WOB to eliminate walks outside4. QMC methods
![Page 194: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/194.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Conclusions and Future Work
Future Work
I Further algorithmic development1. Computation of gradients using existing Markov chains2. Global computation of field variables and their visualization3. Nonlinear BVPs perhaps via branching processes4. Using “Walk-on-the-Boundary" (WOB) techniques
I Geometric Issues1. Computation of the three region model problem2. More complicated surfaces (solvent-excluded and ion-excluded)3. Accuracy issues related to the Van der Waals surface
I Optimize the performance1. Error/bias/variance balancing2. Importance sampling and the outer walks3. WOB to eliminate walks outside4. QMC methods
![Page 195: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/195.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Conclusions and Future Work
Bibliography
[T. Mackoy, R. C. Harris, J. Johnson, M. Mascagni andM. O. Fenley (2011)] Numerical Optimization of aWalk-on-Spheres Solver for the Linear Poisson-BoltzmannEquation Communications in Computational Physics, in thepress.
[M. Fenley, M. Mascagni, J. McClain, A. Silalahi and N. Simonov(2010)] Using Correlated Monte Carlo Sampling for EfficientlySolving the Linearized Poisson-Boltzmann Equation Over aBroad Range of Salt Concentrations Journal of Chemical Theoryand Computation, 6(1): 300–314.
[N. Simonov and M. Mascagni and M. O. Fenley (2007)] MonteCarlo Based Linear Poisson-Boltzmann Approach MakesAccurate Salt-Dependent Solvation Energy Predictions PossibleJournal of Chemical Physics, 187(18), article #185105, 6 pages.
![Page 196: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/196.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Conclusions and Future Work
Bibliography
[T. Mackoy, R. C. Harris, J. Johnson, M. Mascagni andM. O. Fenley (2011)] Numerical Optimization of aWalk-on-Spheres Solver for the Linear Poisson-BoltzmannEquation Communications in Computational Physics, in thepress.
[M. Fenley, M. Mascagni, J. McClain, A. Silalahi and N. Simonov(2010)] Using Correlated Monte Carlo Sampling for EfficientlySolving the Linearized Poisson-Boltzmann Equation Over aBroad Range of Salt Concentrations Journal of Chemical Theoryand Computation, 6(1): 300–314.
[N. Simonov and M. Mascagni and M. O. Fenley (2007)] MonteCarlo Based Linear Poisson-Boltzmann Approach MakesAccurate Salt-Dependent Solvation Energy Predictions PossibleJournal of Chemical Physics, 187(18), article #185105, 6 pages.
![Page 197: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/197.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Conclusions and Future Work
Bibliography
[T. Mackoy, R. C. Harris, J. Johnson, M. Mascagni andM. O. Fenley (2011)] Numerical Optimization of aWalk-on-Spheres Solver for the Linear Poisson-BoltzmannEquation Communications in Computational Physics, in thepress.
[M. Fenley, M. Mascagni, J. McClain, A. Silalahi and N. Simonov(2010)] Using Correlated Monte Carlo Sampling for EfficientlySolving the Linearized Poisson-Boltzmann Equation Over aBroad Range of Salt Concentrations Journal of Chemical Theoryand Computation, 6(1): 300–314.
[N. Simonov and M. Mascagni and M. O. Fenley (2007)] MonteCarlo Based Linear Poisson-Boltzmann Approach MakesAccurate Salt-Dependent Solvation Energy Predictions PossibleJournal of Chemical Physics, 187(18), article #185105, 6 pages.
![Page 198: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/198.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Conclusions and Future Work
Bibliography
[M. Mascagni and N. A. Simonov (2004)] Monte Carlo Methodsfor Calculating Some Physical Properties of Large MoleculesSIAM Journal on Scientific Computing, 26(1): 339–357.
[N. A. Simonov and M. Mascagni (2004)] Random WalkAlgorithms for Estimating Effective Properties of Digitized PorousMedia Monte Carlo Methods and Applications, 10: 599–608.
[M. Mascagni and N. A. Simonov (2004)] The Random Walk onthe Boundary Method for Calculating Capacitance Journal ofComputational Physics, 195: 465–473.
[A. Karaivanova, N. A. Simonov and M. Mascagni(2004)] ParallelQuasirandom Walks on the Boundary Monte Carlo Methods andApplications, 11: 311–320.
![Page 199: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/199.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Conclusions and Future Work
Bibliography
[M. Mascagni and N. A. Simonov (2004)] Monte Carlo Methodsfor Calculating Some Physical Properties of Large MoleculesSIAM Journal on Scientific Computing, 26(1): 339–357.
[N. A. Simonov and M. Mascagni (2004)] Random WalkAlgorithms for Estimating Effective Properties of Digitized PorousMedia Monte Carlo Methods and Applications, 10: 599–608.
[M. Mascagni and N. A. Simonov (2004)] The Random Walk onthe Boundary Method for Calculating Capacitance Journal ofComputational Physics, 195: 465–473.
[A. Karaivanova, N. A. Simonov and M. Mascagni(2004)] ParallelQuasirandom Walks on the Boundary Monte Carlo Methods andApplications, 11: 311–320.
![Page 200: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/200.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Conclusions and Future Work
Bibliography
[M. Mascagni and N. A. Simonov (2004)] Monte Carlo Methodsfor Calculating Some Physical Properties of Large MoleculesSIAM Journal on Scientific Computing, 26(1): 339–357.
[N. A. Simonov and M. Mascagni (2004)] Random WalkAlgorithms for Estimating Effective Properties of Digitized PorousMedia Monte Carlo Methods and Applications, 10: 599–608.
[M. Mascagni and N. A. Simonov (2004)] The Random Walk onthe Boundary Method for Calculating Capacitance Journal ofComputational Physics, 195: 465–473.
[A. Karaivanova, N. A. Simonov and M. Mascagni(2004)] ParallelQuasirandom Walks on the Boundary Monte Carlo Methods andApplications, 11: 311–320.
![Page 201: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/201.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Conclusions and Future Work
Bibliography
[M. Mascagni and N. A. Simonov (2004)] Monte Carlo Methodsfor Calculating Some Physical Properties of Large MoleculesSIAM Journal on Scientific Computing, 26(1): 339–357.
[N. A. Simonov and M. Mascagni (2004)] Random WalkAlgorithms for Estimating Effective Properties of Digitized PorousMedia Monte Carlo Methods and Applications, 10: 599–608.
[M. Mascagni and N. A. Simonov (2004)] The Random Walk onthe Boundary Method for Calculating Capacitance Journal ofComputational Physics, 195: 465–473.
[A. Karaivanova, N. A. Simonov and M. Mascagni(2004)] ParallelQuasirandom Walks on the Boundary Monte Carlo Methods andApplications, 11: 311–320.
![Page 202: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/202.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Conclusions and Future Work
Bibliography
[C.-O. Hwang and M. Mascagni (2003)] Analysis andComparison of Green’s Function First-Passage Algorithms with“Walk on Spheres” Algorithms Mathematics and Computers inSimulation, 63: 605–613.
[C.-O. Hwang and M. Mascagni (2001)] Efficient modified “walkon spheres" algortihm for the linearized Poisson-Boltzmannequation Applied Physics Letters, 78: 787–789.
[C.-O. Hwang, J. A. Given and M. Mascagni (2001)] TheSimulation-Tabulation Method for Classical Diffusion Monte CarloJournal of Computational Physics, 174: 925–946.
[C.-O. Hwang, J. A. Given and M. Mascagni (2000)] On the RapidCalculation of Permeability for Porous Media Using BrownianMotion Paths Physics of Fluids, 12: 1699–1709.
![Page 203: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/203.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Conclusions and Future Work
Bibliography
[C.-O. Hwang and M. Mascagni (2003)] Analysis andComparison of Green’s Function First-Passage Algorithms with“Walk on Spheres” Algorithms Mathematics and Computers inSimulation, 63: 605–613.
[C.-O. Hwang and M. Mascagni (2001)] Efficient modified “walkon spheres" algortihm for the linearized Poisson-Boltzmannequation Applied Physics Letters, 78: 787–789.
[C.-O. Hwang, J. A. Given and M. Mascagni (2001)] TheSimulation-Tabulation Method for Classical Diffusion Monte CarloJournal of Computational Physics, 174: 925–946.
[C.-O. Hwang, J. A. Given and M. Mascagni (2000)] On the RapidCalculation of Permeability for Porous Media Using BrownianMotion Paths Physics of Fluids, 12: 1699–1709.
![Page 204: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/204.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Conclusions and Future Work
Bibliography
[C.-O. Hwang and M. Mascagni (2003)] Analysis andComparison of Green’s Function First-Passage Algorithms with“Walk on Spheres” Algorithms Mathematics and Computers inSimulation, 63: 605–613.
[C.-O. Hwang and M. Mascagni (2001)] Efficient modified “walkon spheres" algortihm for the linearized Poisson-Boltzmannequation Applied Physics Letters, 78: 787–789.
[C.-O. Hwang, J. A. Given and M. Mascagni (2001)] TheSimulation-Tabulation Method for Classical Diffusion Monte CarloJournal of Computational Physics, 174: 925–946.
[C.-O. Hwang, J. A. Given and M. Mascagni (2000)] On the RapidCalculation of Permeability for Porous Media Using BrownianMotion Paths Physics of Fluids, 12: 1699–1709.
![Page 205: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/205.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Conclusions and Future Work
Bibliography
[C.-O. Hwang and M. Mascagni (2003)] Analysis andComparison of Green’s Function First-Passage Algorithms with“Walk on Spheres” Algorithms Mathematics and Computers inSimulation, 63: 605–613.
[C.-O. Hwang and M. Mascagni (2001)] Efficient modified “walkon spheres" algortihm for the linearized Poisson-Boltzmannequation Applied Physics Letters, 78: 787–789.
[C.-O. Hwang, J. A. Given and M. Mascagni (2001)] TheSimulation-Tabulation Method for Classical Diffusion Monte CarloJournal of Computational Physics, 174: 925–946.
[C.-O. Hwang, J. A. Given and M. Mascagni (2000)] On the RapidCalculation of Permeability for Porous Media Using BrownianMotion Paths Physics of Fluids, 12: 1699–1709.
![Page 206: Monte Carlo Methods for Partial Differential Equations: A ... · Monte Carlo Methods for Partial Differential Equations: A Personal Journey Monte Carlo Methods for Partial Differential](https://reader033.vdocument.in/reader033/viewer/2022042923/5f6fea48417d52545919966c/html5/thumbnails/206.jpg)
Monte Carlo Methods for Partial Differential Equations: A Personal Journey
Conclusions and Future Work
c© Michael Mascagni, 2012