killer lab: flow simulation and lead poisoning study james heliotis, computer science carl lutzer,...

19
Killer Lab: Flow Simulation and Lead Poisoning Study James Heliotis, Computer Science Carl Lutzer, Mathematics Rochester Institute of Technology

Upload: dorcas-whitehead

Post on 16-Dec-2015

216 views

Category:

Documents


1 download

TRANSCRIPT

Page 1: Killer Lab: Flow Simulation and Lead Poisoning Study James Heliotis, Computer Science Carl Lutzer, Mathematics Rochester Institute of Technology

Killer Lab:Flow Simulation and

Lead Poisoning Study

James Heliotis, Computer Science

Carl Lutzer, Mathematics

Rochester Institute of Technology

Page 2: Killer Lab: Flow Simulation and Lead Poisoning Study James Heliotis, Computer Science Carl Lutzer, Mathematics Rochester Institute of Technology

October 2005 Killer Examples Workshop 2

RIT’s Organization

Quarter System Mandatory co-op Shared 1st-year curriculum in

computing: Computer Science Computer Engineering Software Engineering

Page 3: Killer Lab: Flow Simulation and Lead Poisoning Study James Heliotis, Computer Science Carl Lutzer, Mathematics Rochester Institute of Technology

October 2005 Killer Examples Workshop 3

Our Introductory Course Curriculum: Year 1 CS1: Computing/Programming intro CS2: "Engineering": inheritance, library

use, GUI, threads, networking, design patterns

CS3: "Computer Science": recursion, time complexity, data structures, design patterns

Page 4: Killer Lab: Flow Simulation and Lead Poisoning Study James Heliotis, Computer Science Carl Lutzer, Mathematics Rochester Institute of Technology

October 2005 Killer Examples Workshop 4

Our Introductory Course Curriculum: Year 2 CS4: UML, Design, C++, Team projects SE1: Development Methodology /

Process, Larger Project, design patterns

Page 5: Killer Lab: Flow Simulation and Lead Poisoning Study James Heliotis, Computer Science Carl Lutzer, Mathematics Rochester Institute of Technology

October 2005 Killer Examples Workshop 5

Learning Communities

Pilot program at RIT 2004-05 Small (~20) group of students stick

together as a group for several of their courses.

Our L.C.: Writing/Lit.(2Q), Calculus(3Q), CS(3Q), First Year Enrichment(2Q)

Page 6: Killer Lab: Flow Simulation and Lead Poisoning Study James Heliotis, Computer Science Carl Lutzer, Mathematics Rochester Institute of Technology

October 2005 Killer Examples Workshop 6

Our LC

Not a lot of shared curriculum! Coordinated by:

Ensuring multiple tests did not occur on the same date

Discussing issues with particular students

Page 7: Killer Lab: Flow Simulation and Lead Poisoning Study James Heliotis, Computer Science Carl Lutzer, Mathematics Rochester Institute of Technology

October 2005 Killer Examples Workshop 7

Things We Did Together

All profs attended showing of movie to be discussed in literature course.

Math prof donated initial chapters of fiction novel in progress to be critiqued.

CS prof participated in writing an ode to his computer along with his students.

Shared project in math and CS courses…

Page 8: Killer Lab: Flow Simulation and Lead Poisoning Study James Heliotis, Computer Science Carl Lutzer, Mathematics Rochester Institute of Technology

October 2005 Killer Examples Workshop 8

Calculus Project

…One of the principle questions in the community at large is whether people can recover from prolonged exposure to toxic elements. This will be quite impossible unless the body is able to fully flush the toxin from the system.

Suppose a community is exposed to high quantities of lead. The lead is absorbed into the body at a rate of 49.3 micrograms per day and is transported to bones, tissues and organs via the blood.

…lead is transferred between blood, bone and tissue at a rate that is proportional to the amount present.

Page 9: Killer Lab: Flow Simulation and Lead Poisoning Study James Heliotis, Computer Science Carl Lutzer, Mathematics Rochester Institute of Technology

October 2005 Killer Examples Workshop 9

Given Transfer Rates

dsa

dt kiasi

ia

kaisa

ia

Page 10: Killer Lab: Flow Simulation and Lead Poisoning Study James Heliotis, Computer Science Carl Lutzer, Mathematics Rochester Institute of Technology

October 2005 Killer Examples Workshop 10

The Problem Domain

Flow of material through conduits into and out of reservoirs

ingestionBlood

BoneTissues

sweat glands

renal system

Page 11: Killer Lab: Flow Simulation and Lead Poisoning Study James Heliotis, Computer Science Carl Lutzer, Mathematics Rochester Institute of Technology

October 2005 Killer Examples Workshop 11

The Analysis Pattern

A network of nodes A node represents a reservoir of material. Each edge is a conduit; its weight is the k ij

transfer rate.

Nodes come in three variations: inputinput internalinternal outputoutput

Page 12: Killer Lab: Flow Simulation and Lead Poisoning Study James Heliotis, Computer Science Carl Lutzer, Mathematics Rochester Institute of Technology

October 2005 Killer Examples Workshop 12

How the Project is Structured

Students are assigned the project in their Calculus class and begin to solve it.

Students are assigned a graph implementation lab in CS.

Students are assigned the "blood" project in CS.

Page 13: Killer Lab: Flow Simulation and Lead Poisoning Study James Heliotis, Computer Science Carl Lutzer, Mathematics Rochester Institute of Technology

October 2005 Killer Examples Workshop 13

What It Teaches

Reinforces graph theory. Graph implementation trade-offs. Graph traversal algorithms. Analysis pattern: Conduits & Reservoirs The Two-Phase Discrete Simulation

design pattern: a specialized application of the Template Method

Page 14: Killer Lab: Flow Simulation and Lead Poisoning Study James Heliotis, Computer Science Carl Lutzer, Mathematics Rochester Institute of Technology

October 2005 Killer Examples Workshop 14

2-Phase Discrete Simulation

For each node n in the graph, do: Compute n's new value based on edges'

and neighboring nodes' values. Store n's new value in a temporary area.

For each node n in the graph, do: Reveal the new value of n as the true

value of n.

Page 15: Killer Lab: Flow Simulation and Lead Poisoning Study James Heliotis, Computer Science Carl Lutzer, Mathematics Rochester Institute of Technology

October 2005 Killer Examples Workshop 15

Advantages of a Collaborative Project Students get more exposure to the problem.

better comprehension better discussions

Learn connection between mathematics and computer science, and the differences in approaches.

"Answers" from one project can be checked against those from the other.

Page 16: Killer Lab: Flow Simulation and Lead Poisoning Study James Heliotis, Computer Science Carl Lutzer, Mathematics Rochester Institute of Technology

October 2005 Killer Examples Workshop 16

Advantages of a Project in an Application Area Students can relate abstract CS ideas

to real-world problems. Students learn the challenge in

designing the API for a graph class. Operations needed are heavily dependent

on application. Needed operations must be made efficient.

Page 17: Killer Lab: Flow Simulation and Lead Poisoning Study James Heliotis, Computer Science Carl Lutzer, Mathematics Rochester Institute of Technology

October 2005 Killer Examples Workshop 17

Possible Additions

Add a follow up to that uses the same template method for a different purpose, e.g. digital circuit simulation neural network …

Page 18: Killer Lab: Flow Simulation and Lead Poisoning Study James Heliotis, Computer Science Carl Lutzer, Mathematics Rochester Institute of Technology

October 2005 Killer Examples Workshop 18

More Specific Ideas

The Geneva NY Splash Park Incident Directed-Graph Epidemiological Models

of Computer Viruses J. Kephart, S. White, IBM TJWRC

Modeling Functions for Evaluation and Differentiation Just polynomials General functions w/ inheritance, genericity

Composite pattern

Page 19: Killer Lab: Flow Simulation and Lead Poisoning Study James Heliotis, Computer Science Carl Lutzer, Mathematics Rochester Institute of Technology

October 2005 Killer Examples Workshop 19

Where to find it

http://www.cs.rit.edu/~vcss233/Projects/newproj02/writeup.html

My email: [email protected]@rit.edu