Introduction to Agent-Based Modeling

EES 4760/5760

Agent-Based & Individual-Based Computational Modeling

Jonathan Gilligan

Class #1: Tues. January 10 2018

Who Are You?

Who Are You?

  1. Who are you? (Name, year, major)
  2. Computational skills (if any)
    • programming, statistical analysis, …
  3. What do you want to get from this class?
  4. Ask me a question about computational modeling
  5. Something interesting about yourself

Getting Started

Getting Started

For Thursday:

  • Download and install NetLogo on your computer.
    • URL in syllabus and assignment sheet
  • Set up Box account
    • Details in syllabus and assignment sheet
    • https://vanderbilt.box.com
    • Make folder for this class with your last name:
      • lastname_EES_4760 or lastname_EES_5760
      • Share it with me as Editor
      • Homework goes in subfolders:
        • HW_1, HW_2, …

Agent-Based Modeling

Agent-Based Modeling

  • Simulate individuals:
    • Autonomous
    • Heterogeneous
    • Quasi-local
    • Bounded rationality
  • Simulate environment
  • Emphasize simplicity, minimal assumptions
  • Emergence: Large-scale phenomena arise from small-scale individual interactions
    • Interesting when large-scale is not easily predicted from small-scale

Simple Experiments

  • Play with economics
    • Simple agents trade with each other
    • Confirm 1st welfare theorem:

      Trading leads to Pareto equilibrium

    • Find conditions for satisfying theorem:
      • Not necessary for traders to be completely rational
        • How much rationality do you need?
      • Equilibration can be slow
      • Time-varying preferences can prevent equilibration
  • Dynamics of agent-based models connect to nonlinear dynamics and chaos

Economics of Cooperation

Game Theory

  • Prisoner’s Dilemma Game:

    A \ B B Cooperates B Defects
    A Cooperates (3,3) (0,4)
    A Defects (4,0) (1,1)
  • Nash Equilibrium:
    • No matter what player A does, player B is better off defecting
    • No matter what player B does, player A is better off defecting
    • End result: Both players end up worse off than if they had both cooperated.

Iterated Prisoner’s Dilemma

  • R. Axelrod (1981)
  • Tournament of algorithms
  • Winner: “tit-for-tat”
  • Evolutionary Game Theory:
    • Basic principles of good strategies:
      • Be nice
      • Be provocable
      • Don’t be too envious
      • Don’t be too clever
  • Nay & Gilligan (2015)
    • Real-world strategies involve randomness, unpredictability

Artificial Anasazi

Example: Artificial Anasazi

Axtell, Dean, Epstein, et al.

Long House Valley

Long House Valley (flourished ca. 1800 BCE–1300 CE)

Modeling Environment

Artificial Anasazi Simulations

Constructing model

  • Paleoclimate:
    • Assess different kinds of soil
    • Assess tree rings, pollen, etc.
    • Reconstruct drought severity index
  • Society:
    • Archaeology gives #, location of households
  • Make assumptions about:
    • # people per household,
    • Agriculture,
  • Devise rules for behavior:
    • Marriage, reproduction, migration, …
  • Simulate years 800–1300

Results

Artificial Ansazi model results

Comparison

Simulated Historical
Artificial Ansazi simulation Artificial Ansazi simulation

Improvements

  • Make agents heterogeneous
  • Fit parameters to historical data

Results

Artificial Ansazi improved model results