Principles of agent-based modeling Computational Modeling of Social - PowerPoint PPT Presentation

Introduction to Computational Modeling of Social Systems Introduction to Principles of agent-based modeling Computational Modeling of Social Systems Prof. Lars-Erik Cederman ETH - Center for Comparative and International Studies (CIS) Seilergraben 49, Room G.2, lcederman@ethz.ch Christa Deiwiks, CIS Room E.3, deiwiks@icr.gess.ethz.ch http://www.icr.ethz.ch/teaching/compmodels Week 3

Today’s agenda 2 • Prehistory • Other types of models • Principles of agent based modeling • Categories of ABM models • The pros and cons of ABM

Historical Lineages of ABM 3 Source: Nigel Gilbert

Von Neumann’s theory of cellular automata 4 • Cellular automata are discrete dynamical systems that model complex behavior based on simple, local rules animating cells on a lattice Invented by John von Neumann

Game of Life 5 • First practical CA invented by John Conway in the late 1960s • Later popularized by Martin Gardner John Conway Simple rules: •A dead cell with 3 live neighbors comes to life •A live cell with 2 or 3 neighbors stays alive Stephen Wolfram •Otherwise the cell dies Expert on CAs http://www.ibiblio.org/lifepatterns/

Four types of models 6 Modeling language : Analytical Deductive Computational focus : 1. Analytical 2. Macro- Systemic macro models simulation variables 3. Rational Micro- 4. Agent-based choice modeling mechanisms

1. Analytical macro models 7 • Equilibrium conditions or systemic variables traced in time • Closed-form, and often based on differential equations • Examples: macro economics and traditional systems theory

2. Macro simulation 8 • Dynamic systems, tracing macro variables over time • Based on simulation • Systems theory and Global Modeling Jay Forrester, MIT

3. Rational choice modeling 9 • Individualist reaction to macro approaches • Decision theory and game theory • Analytical equilibrium solutions • Used in micro-economics and spreading to other social sciences

4. Agent-based modeling 10 • ABM is a computational methodology that allows the analyst to create, analyze, and experiment with, artificial worlds populated by agents that interact in non-trivial ways • Bottom-up • Computational • Builds on CAs and DAI

Complex Adaptive Systems 11 A CAS is a network exhibiting aggregate properties that emerge from primarily local interaction among many, typically heterogeneous agents mutually constituting their own environment. � Emergent properties � Large numbers of diverse agents � Local and/or selective interaction � Adaptation through selection � Endogenous, non-parametric environment

Microeconomics � ABM 12 � Analytical Synthetic approach � Equilibrium Non-equilibrium theory � Nomothetic Generative method Variable-based � Configurative ontology

Analytical � Synthetic approach 13 • Hope to solve problems through strategy of “divide and conquer” • Need to make ceteris paribus assumption • But in complex systems this assumption breaks down • Herbert Simon: Complex systems are composed of large numbers of parts that interact in a non-linear fashion • Need to study interactions explicitly

Equilibrium � Non-equilibrium theory 14 • Standard assumption in the social sciences: “efficient” history • But contingency and positive feedback undermine this perspective • Complexity theory and non- equilibrium physics • Statistical regularities at the macro level despite micro-level contingency Example: Avalanches in rice pile

� Nomothetic Generative method 15 • Search for causal regularities • Hempel’s “covering laws” • But what to do with complex social systems that have few counterparts? • Scientific realists explain complex patterns by deriving the mechanisms that generate them • Axelrod: “third way of doing science” • Epstein: “if you can’t grow it, you haven’t explained it!”

Variable-based � Configurative ontology 16 • Conventional models are variable- based • Social entities are assumed implicitly • But variables say little about social forms • A social form is a configuration of social interactions and actors together with the structures in which they are embedded • ABM good at endogenizing interactions and actors • Object-orientation is well suited to capture agents

Emergent social forms 17 1. Interaction patterns 2. Property configurations 3. Dynamic networks 4. Actor structures

1. Emergent interaction patterns 18 • Models of “emergent order” producing configurations • Axelrod (1984, chap. 8): actor actor actor “The structure of actor actor actor cooperation” actor actor actor

2. Emergent property configurations 19 • Models of “emergent structure” constituted as property configruations • Example: Schelling’s segregation model; Carley 1991; Axelrod 1997 • See Macy 2002 for further actor actor actor references actor actor actor actor actor actor actor actor actor actor actor actor actor actor actor

3. Emergent dynamic networks 20 • Most computational frequency models treat networks as exogenous d - α • Recent exceptions: – Albert and Barabási’s scale- degree d free networks – Economics and evolutionary game theory: e.g. Skyrms and Pemantle

4. Emergent actor structures 21 • Computational models normally assume the actors to be given • Exceptions: – Axelrod’s model of new political actors – Axtell’s firm-size model – Geopolitical models in the Bremer & Mihalka tradition • Emergence?

Why use agent-based modeling? 22 ABM helps fill the gap between formal but restrictive models and wide-ranging but imprecise qualitative frameworks � Contingency and counterfactuals � Mechanisms in space and time � Intangible concepts

The limits of ABM? 23 ad hoc assumptions? failure to yield predictions? fragility of results? lack of cumulation? Horgan, John. 1995. "From Complexity to Perplexity." Scientific American 272: 104-9. Binmore, Ken. 1998. Review of The Complexity of Cooperation . JASSS 1: http://jasss.soc.surrey.ac.uk/1/1/review1.html See also references in Rosser (recommended reading)