agent-based models of animal behaviour - universiteit utrecht · agent-based models = agent-based...
TRANSCRIPT
Agent-based models
of animal behaviour
From micro to macro
and back.
Ellen Evers, Behavioural Biology, University Utrecht.
Outline
... - genome – cell – organism – group – population – ecosystem - ...
� Organization, structure, patterns(spatial, temporal, social, ...)
� Behaviour of animal groups(coordination)
� Self-organization(complexity arises from simple rules)
� Tool to understand and study: Agent-based Models (ABM)
Flocking
� Examples real:
� Starlings, ants (movie), fish
Collective motion
� Bird flocks (starlings), insect swarms (flies), mammal herds (wildebeest), fish schools (herring)
Fish schooling
School level:
� Locomotion:Simultaneous change of direction
� Defense:Flash expansion, fountain effect, aggregation
No disorganized crowd, but coordinated behaviour between members of the group!
HOW to get structure, patterns, organization in a system?
Organization
Leader:
� Well informed (top-down)
� Provides everyone with instructions
� No! Fish change position within school
Blueprint, Recipe, Template:
� Plan or procedure
� Fixed, not very flexible
� No! Unlikely in big groups
Fish schooling
Individual level:
� Vision (location of neighbours)
� Lateral line (orientation of neighbours)
Coordination only with nearest neighbours, whole swarm locomotion through:
SELF-ORGANIZATION!
Self-Organization
... process in which a pattern at the global level of a system emerges solely from numerous interactions among the lower-level components of the system. (Camezine et al, 2001)
� No external, centralized control (no leader)
� Local information, simple rules (no global plan, overview)
� ORGANIZATION! (bottom-up)
FISH SCHOOLING? COLLECTIVE MOTION?
Fish schooling
Self-Organization Hypothesis:
Few simple rules in response to local information from neighbouring fish generate schooling patterns without directly coding for them. (Huth & Wissel, 1992)
Schooling rules:
1. Cohesion 2. Separation 3. Alignment
Fish schooling
Self-Organization Hypothesis:
Few simple rules in response to local information from neighbouring fish generate schooling patterns without directly coding for them. (Huth & Wissel, 1992)
To test hypothesis and understand mechanism:
Models?!
Agent-Based Models (ABM)
Models
What is a model?
Models
Models
Models
Models
Models
Definition:
Representation of a system of entities, phenomena, or processes
to simplify, visualize, simulate, manipulate and gain intuition about the entity, phenomenon or process being represented.
Simplified Realistic Understandable Predictive Explaining Precise
Simplification versus Realism
www.wikipedia.org
Collective motion
Model rules:
Craig Reynolds 1987: bird flocking (BOIDS)
Huth & Wissel 1992: fish schooling
Hooper, 1999: predator avoidance (cool school)
1. Cohesion 2. Separation 3. Alignment
Collective motion
Conclusions from the model:
� Simple rules can account for complex pattern
� No leader, no external cue nessecary
� Schools could be self-organizing
� No proof of real mechanism
� Proven: possible explanation works
2nd part
Scientific models
Spatial CA nonlineardynamical
ordinary differential equationsdeterministic stochastic IBM
monte carlo simulation IOM
partial differential equationsagent-based models
complex systemsstatic probabalistic linear
ABM
difference equation ODE PDE
cellular automata
AGENT-BASED MODELS(ABM)
ORDINARYDIFFERENTIAL EQUATIONS
(ODE)
-
Predator-prey system
sheep wolves
+
sheep sheep wolves
wolves
a) data: b) model:
-+
-
sheep wolves+
+ -
ODE:
� Population densities
ABM:
� Discrete indiviuals
sheep wolves
-
sheep wolves+
+ -
ODE:
� Equations
ABM:
� Rules (if-then, loop, scheduling)
dS/dt = bS – pSWdW/dt = pSW - dW
Sheep = + birth*Sheep– pred.*Sheep*Wolves
Wolves = + pred.*Sheep*Wolves– death*Wolves
SHEEP:
If I meet wolve: die!With chance = birthrate: Reproduce!
WOLVES:
If I meet sheep: energy +1If energy (from sheep) = 0: die!With chance = birthrate: Reproduce!
-
sheep wolves+
+ -
ODE:
� Deterministic
ABM:
� Stochasticity(pseudo-random numbers)
... to represent
variability of processes
that cannot or need notto be modelled deterministically
sheep wolves
After Grimm 2005
-
sheep wolves+
+ -
ODE:
� Non-spatial
� Homogenity
� Well mixed
ABM:
� Explicitly spatial
� Heterogenity
� Local interactions(bottom-up, adaptive)
� Patterns
-
sheep wolves+
+ -
ODE:
� Homogeneous population
ABM:
� Individual variation(age, colour, knowledge, spatial position, size, sex, ...)
-
sheep wolves+
+ -
Agent-based Models
= Agent-based Models (ABM)
= Individual-based Models (IBM)
= Individual-oriented Models (IOM)
... describe behaviour, variability and
interactions of autonomous individuals.
After Grimm 2005
Collective foraging
� Ants: Trail formation, transporter recruitment, efficient path finding, PHEROMONES!
Collective foraging
Collective decision of colony:
� Selection of richest of two foods
� Selection of shortest path
� Hyp 1: directed by leader (queen?)
� Hyp 2: Self-organization
� Global pattern emergesfrom simple rules and local information
Collective foraging
Foraging rules:
� Execute random walkor follow strongest pheromone trail
� Food found: walk home & leave pheromone trace
Resnick, M. (1994)
Collective foraging
Self-organization:
� Simple rules
� No global knowledge, local information
� Self-reinforcing phermone trails
� Feedback via environment
Emergence of accurate (not perfect!) collective decision:efficient path selection.
Self-organization
Mechanism:
� Information from neighbours
� Information from local environment (stigmergy)
� Feedback from emerging structure(control, signal)
Emergent properties
... arise out of interactions between lower-level entities, yet are novel and irreducible to them.
(Stanford Encyclopedia of Philosophy)
Emergent property:
Order, organization,
pattern on group level
Entities:
Individuals without
global knowledge,
plan or intent
Emergence
� Self-organization
� Complex Systems� physical and chemical systems (liquid)
� biological systems (cell, brain)
� social systems and organizations (stock market, social structures)
Complex system patterns often
emerge through self-organization.
Complex Systems
Multiple levels of organization!
� Individual < subgroup < population
� Interconnected, interdependent
� Higher level affects lower one
� Higher level has “behaviour” on its own
� Emergent property: Substrate to evolution
3rd part
ABM in science?
Criticism:
� Too complex to be understood
� Impossible to include everything in a model
� Too many parameters unknown
� Hard to test
� Parameters can tweak everything
� “Gaming”
After Grimm 2005
Modelling cycle
After Grimm 2005
Modelling cycle
After Grimm 2005
Formulate the question:
� Filter for essential processes
� Not model per se,but for specific purpose
Modelling cycle
After Grimm 2005
Assemble hypotheses:
� Starting point: conceptual model (drawing)
� Based on empirical data,experience, theory, gut feeling
� What do we expect to see?PATTERNS (at multiple levels)
Modelling cycle
After Grimm 2005
Chose model structure:
� Which entities?
� Which processes?
� How to model/represent processes?
Modelling cycle
After Grimm 2005
Implement model:
� Write program code(if-then rules, loops)
Modelling cycle
After Grimm 2005
Analyze the model I:
� Controlled simulation experiments(visual debugging, process understanding,extreme tests)
� Design and analysis just as with “real” experiments(statistics on program generated data)
Modelling cycle
After Grimm 2005
Analyze the model II:
� Much freedom and control in experimental setup(upscale numbers and time, hypothetical thought experiments)
� Can the model reproduce PATTERNS?(at multiple levels!!!)
Group
Individual
Group
Individual
Group
Individual
Group
Individual
Group
Individual
Group
Individual
Group
Individual
Group
Individual
Same parametersetting:
Changing parameter
setting:
Modelling cycle
After Grimm 2005
Analyze the model IIIa:
� Change parameter/rule: Effect on system behaviour?(knockout, sufficient & neccessary)
Modelling cycle
Analyze the model IIIb:
� Multiple runs, same parameter set(account for stochasticity)
After Grimm 2005
Group
Individual
Group
Individual
Group
Individual
Group
Individual
Same parametersetting:
Modelling cycle
After Grimm 2005
Communicate the model:
� Materials & Methods
� Reproducability of results
ABM – a scientific method
Model validation:
� New hypotheses from the model
� Not built into the model beforehand
� Can be tested in “real” system
After Grimm 2005
Primate social cognition
� Social structure: dominance hierachy
� Spatial structure: centrality of dominants
Primate social cognition
� DOMWORLD (Hemelrijk 1999, Hogeweg1988)
� Simple rules > complex behaviour
1. Grouping rules:
Max view
Near view
Personal space
Primate social cognition
� Dom-value: rank(initially all the same)
2. Interaction rules:
� Prediction(mental simulation)
� Dominance contest:� Win chance ~ rank
� Loser flees, winner chases
� Dom-value update according to expected outcome(winner-loser effect)
Primate social cognition
� Dominance hierarchy (differentiation from initially equal individuals)(individual variation!!!)
� Emergence of spatial structure(centrality of dominant individuals as unintentional side-effect)
� Stabilizing feedback from spatial structure on social structure:� neighbors of similar rank
� no unexpected outcomes
� no big updates of dom-value
Hemelrijk, 1998
Primate social cognition
INSIGHTS:
� No centripetal instinct necessary for spatial structure� From side effect to evolutionary substrate� Emerging spatial structure feeds back on social structure
ABM internship: http://bio.uu.nl/behaviour/people/Evers/index.html
Evolution in silico
Evolution ingredients:
� Mutations:� Inheritence of traits (to offspring)
� Chance of changing traits
� Selection:� Define fitness measure
� Explicit selection(once in a while: take out 100 best and reproduce)
� Reproduction/death(dependent on fitness measure)
Conclusions
Biological systems:
� Organisation / patterns
Several possibilities:
� Central, fixed control (leader, blueprint, ...)
� Self-organisation� Interactions between individuals
� Emergence of pattern on group level
Conclusions
ABM: useful tool to study / understand
� Multiple connected levels (individual, group)
� Behavioural rules
� Explicit spatiality:
� Locality
� Heterogenous space
� Individual variation (hierarchy, evolution)
Heraklit: You never enter the same river twice.
Conclusions
Models are useful:
� Development: Explicit formulation of processes
� Simplification: Identify essential (sufficient & nessecary) processes
� Scientific research tool:
� Controlled experiments (knockout, change parameter/rule)
� Series of runs (variation)
� Statistics on generated data
� Material & methods
� Reproduced results: validation
Group
Individual
Group
Individual
Group
Individual
Conclusions
Models advantages:
� Freedom in experimental setup
� Little time, money, animals, facilities needed
� No ethical objections
� Generate new hypotheses
� Reveal general and unexpected properties of a system
� Help understanding a system
� Change way of thinking about a system
Proust: The real voyage of discovery lies not in
finding new landscapes, but in having new eyes.
Sources
Modelling, generalModelling, generalModelling, generalModelling, general
J. J. BrysonJ. J. BrysonJ. J. BrysonJ. J. Bryson, Y. Ando, and H. Lehmann, “Agent-based modelling as scientific method: a case study analysing primate social behaviour,”Philosophical Transactions of the Royal Society B: Biological Sciences 362, no. 1485 (2007): 1685-1698.
D. L. DeAngelisD. L. DeAngelisD. L. DeAngelisD. L. DeAngelis and W. M. Mooij, Individual-based modeling of ecological and evolutionary processes. Annual Reviews in Ecology, Evolution and Systematics 36: 147-168.
V. GrimmV. GrimmV. GrimmV. Grimm, “Ten years of individual-based modelling in ecology: what have we learned and what could we learn in the future?,” Ecological Modelling 115, no. 2 (1999): 129-148.
V. GrimmV. GrimmV. GrimmV. Grimm and S. F. Railsback, Individual-based Modeling and Ecology (Princeton University Press, 2005). V. GrimmV. GrimmV. GrimmV. Grimm et al., “Pattern-Oriented Modeling of Agent-Based Complex Systems: Lessons from Ecology,” Science 310, no. 5750 (November 11,
2005): 987-99. J. W. HaefnerJ. W. HaefnerJ. W. HaefnerJ. W. Haefner, Modeling Biological Systems: Principles and Applications, 2nd ed. (Springer, 2005). C. K. HemelrijkC. K. HemelrijkC. K. HemelrijkC. K. Hemelrijk, “Understanding Social Behaviour with the Help of Complexity Science (Invited Article),” Ethology 108, no. 8 (2002): 655-671. P. HogewegP. HogewegP. HogewegP. Hogeweg, “MIRROR beyond MIRROR, Puddles of LIFE,” Artificial Life (1989): 297–316. T. J. PrescottT. J. PrescottT. J. PrescottT. J. Prescott, J. J. Bryson, and A. K. Seth, “Introduction. Modelling natural action selection,” Philosophical Transactions of the Royal Society
B: Biological Sciences 362, no. 1485 (2007): 1521-1529.
http://www.swarm.org/index.php/Main_Pagehttp://ccl.northwestern.edu/netlogo/http://www.abmsystemsbiology.info
Complexity, Emergence and SelfComplexity, Emergence and SelfComplexity, Emergence and SelfComplexity, Emergence and Self----organization, generalorganization, generalorganization, generalorganization, general
http://www.santafe.edu/http://www.vs.uni-kassel.de/systems/index.php/Emergencehttp://www.pbs.org/wgbh/nova/sciencenow/3410/03.htmlhttp://www.cna.org/isaac/Glossb.htm
Biological examples of emergence and selfBiological examples of emergence and selfBiological examples of emergence and selfBiological examples of emergence and self----organizationorganizationorganizationorganization
J. BuckJ. BuckJ. BuckJ. Buck, “Synchronous Rhythmic Flashing of Fireflies. II.,” The Quarterly Review of Biology 63, no. 3 (1988): 265. S. CamazineS. CamazineS. CamazineS. Camazine et al., Self-Organization in Biological Systems: (Princeton University Press, 2003). http://mute-net.sourceforge.net/howAnts.shtml
Sources
AgentAgentAgentAgent----based models of animal behaviourbased models of animal behaviourbased models of animal behaviourbased models of animal behaviour
R. AxelrodR. AxelrodR. AxelrodR. Axelrod, The Complexity of Cooperation: Agent-Based Models of Competition and Collaboration (Princeton University Press, 1997). R. AxelrodR. AxelrodR. AxelrodR. Axelrod and W. D. Hamilton, “The evolution of cooperation,” Science 211, no. 4489 (1981): 1390-1396. H. de VriesH. de VriesH. de VriesH. de Vries and J. C. Biesmeijer, “Modelling collective foraging by means of individual behaviour rules in honey-bees,” Behavioral Ecology and
Sociobiology 44, no. 2 (1998): 109-124. H. de VriesH. de VriesH. de VriesH. de Vries and J. C. Biesmeijer, “Self-organization in collective honeybee foraging: emergence of symmetry breaking, cross inhibition and
equal harvest-rate distribution,” Behavioral Ecology and Sociobiology 51, no. 6 (2002): 557-569. C. K. HemelrijkC. K. HemelrijkC. K. HemelrijkC. K. Hemelrijk, “Towards the integration of social dominance and spatial structure,” Animal Behaviour 59, no. 5 (May 2000): 1035-1048,
doi:10.1006/anbe.2000.1400. C. K. HemelrijkC. K. HemelrijkC. K. HemelrijkC. K. Hemelrijk, “Social phenomena emerging by self-organisation in a competitive, virtual world (‘DomWorld’),” 1st International Conference
of the European Society for Social Simulation (ESSA 2003), Groningen, September (2003). P. HogewegP. HogewegP. HogewegP. Hogeweg, “MIRROR beyond MIRROR, Puddles of LIFE,” Artificial Life (1989): 297–316. P. HogewegP. HogewegP. HogewegP. Hogeweg and B. Hesper, “The ontogeny of the interaction structure in bumble bee colonies: A MIRROR model,” Behavioral Ecology and
Sociobiology 12, no. 4 (1983): 271-283. C. HonkC. HonkC. HonkC. Honk and P. Hogeweg, “The ontogeny of the social structure in a captive Bombus terrestris colony,” Behavioral Ecology and Sociobiology
9, no. 2 (1981): 111-119. H. LehmannH. LehmannH. LehmannH. Lehmann, J. Wang, and J. J. Bryson, “Tolerance and Sexual Attraction in Despotic Societies: A Replication and Analysis of Hemelrijk
(2002),” Modelling Natural Action Selecton: Proceedings of an International Workshop, Edinburgh. The Society for the Study of ArtificialIntelligence and the Simulation of Behaviour (AISB) (2005).
S. C. PrattS. C. PrattS. C. PrattS. C. Pratt et al., “An agent-based model of collective nest choice by the ant Temnothorax albipennis,” Animal Behaviour 70, no. 5 (2005): 1023-1036.
M. ResnickM. ResnickM. ResnickM. Resnick, Turtles, Termites, and Traffic Jams: Explorations in Massively Parallel Microworlds (MIT Press, 1994). C. W. ReynoldsC. W. ReynoldsC. W. ReynoldsC. W. Reynolds, “Flocks, Herds, and Schools: A Distributed Behavioral Model,” Computer Graphics 21, no. 4 (1987): 25-34. W. I. SellersW. I. SellersW. I. SellersW. I. Sellers, R. A. Hill, and B. S. Logan, “An agent-based model of group decision making in baboons,” Philosophical Transactions of the
Royal Society B: Biological Sciences 362, no. 1485 (2007): 1699-1710. D. J. van der PostD. J. van der PostD. J. van der PostD. J. van der Post and P. Hogeweg, “Resource distributions and diet development by trial-and-error learning,” Behavioral Ecology and
Sociobiology 61, no. 1 (2006): 65-80. D. J. van der PostD. J. van der PostD. J. van der PostD. J. van der Post and P. Hogeweg, “Diet traditions and cumulative cultural processes as side-effects of grouping,” Animal Behaviour 75, no. 1
(2008): 133-144.
http://www.red3d.com/cwr/boids/