Agent-Based Modelling in NetLogo (ABM-NetLogo)

Lab Sessions

Below are listed the provisional planned lab sessions. However, these are subject to change. The aim is to respond to student interests and also to support them while working on their own projects.

The slides briefly introduce the topic. Work through the exercise sheets and complete the tasks. The videos cover some of the material in the exercise sheets if you get stuck.

1. First NetLogo session

What are ABM? What is NetLogo? The basic NetLogo environment; Using the command center to directly execute commands; Turtles, Patches and Links; Agentsets and lists; Creating buttons and sliders; Using built-in help: manual and dictionary; Models library.

Slides: Intro. slides (pdf).
Exercise sheet: Exercise 1 - a first netlogo session (pdf).
Video: Quck start video to NetLogo (youtube).
Link: 6-page Quick guide to the language (pdf) by Luis R. Izquierdo. For those with programming experience in other languages.

2. First NetLogo program

The structure of a NetLogo program; Setup and Go procedures; Birth and death of turtles; Adding outputs to the interface screen;  Turtle variables and global variables.

Slides: First netlogo program (pdf).
Exercise sheet: Exercise 2 - a first netlogo program (pdf).
Video: part 1, part 2, part 3 (youtube).
Code: Final program after tasks completed (netlogo)

3. Segregation model

Brief background and motivation of the model. Step-by-step explanation of the code of the Schelling segregation model in NetLogo; Modify the model to incorporate more colours and a different movement rule. Implement a procedure to count the number of distinct clusters.

Slides: Segregation model (pdf).
Exercise sheet: Exercise 3 - segregation model (pdf).
Video: part 1, part 2, part 3 (youtube).
Paper: Schelling, Thomas C. (1971) Dynamic Models of Segregation. Journal of Mathematical Sociology 1:143-186. (pdf)
NetLogo Models Library: Sample Models > Social Science > Segregation
Code: Final program after tasks completed (netlogo)

4. Predation model

Step-by-step explantion of the code of the Wolf Sheep predation model. This model demonstrates multiple dynamically interacting turtle breeds and was developed for education purposes. Modify the model to incorporate another preditor who feeds on wolves.

Slides: Predation model (pdf)
Exercise sheet: Exercise 4 - wolf sheep predation model (pdf).
Video: part 1, part 2, part 3, part 4 (youtube)
Link: Tutorial #1 in the NetLogo manual uses this model: https://ccl.northwestern.edu/netlogo/docs/tutorial1.html

NetLogo Models Library: Sample Models > Biology > Wolf Sheep Predation
Code: humans eat wolves (netlogo), humans and wolves fight v1 (netlogo), humans and wolves fight v2 (netlogo)

Paper: Wilensky, U. & Reisman, K. (2006). Thinking like a Wolf, a Sheep or a Firefly: Learning Biology through Constructing and Testing Computational Theories – an Embodied Modeling Approach. Cognition & Instruction, 24(2), pp. 171-209. (pdf)

5 & 6. Networks

Using links to create networks. Create random and other common network topologies. Preferential attachment and small-world construction algorithms. Clustering coefficent and average path length measures. Determining if a network is connected. Virus on a network model. Modify it to incorporate pref. attach. network. Implement turtles that move on a network.

Slides: Networks (pdf)
Exercise sheet: Exercise 5 - preferential attachment model (pdf). (note: this exercise will not be covered in detail in the lab)
Video: part 1 (undirected links), part 2 (directed links), part 3 (network examples program), part 4 (pref attach), part 5 (small world), part 6 (connected), part 7 (clustering coeff.)

Code: network examples program (netlogo), random network (netlogo), pref. attach and small-world (netlogo), virus on a network with extra networks (netlogo), pref. attach model with Exercise 5 tasks competed (netlogo)
NetLogo Models Library: Sample Models > Networks > Preferential Attatchment | Small Words | Virus on a Network | Team Assembly
NetLogo Models Library: Sample Models > Code Examples > Link-walking turtles example | Lattice-walking turtles example

Link: The NetLogo network extension supports many useful network functions: http://ccl.northwestern.edu/netlogo/docs/nw.html

Paper: Albert-László Barabási & Reka Albert. Emergence of Scaling in Random Networks, Science, Vol 286, Issue 5439, 15 October 1999, pages 509-512. (pdf)
Paper: DJ Watts and SH Strogatz. Collective dynamics of ‘small-world’ networks, Nature, 393:440-442 (1998). (pdf)
Paper: R Guimera, B Uzzi, J Spiro, L Amaral (2005) Team Assembly Mechanisms Determine Collaboration Network Structure and Team Performance. Science, 308(5722), p697-702. (pdf)

7. Lists and the El Farol Bar model

Describe some list primatives. Implement simple model in which agents store and update lists. Introduce Brian Arthur's El Farol Bar model. Describe NetLogo model implementation step-by-step. This model demonstrates the use of lists to compactly represent simple learning in turtles. Modify to include new output measures, correct the learning function and give agents the ability to adapt strategies.

Slides: Lists and El Farol model (pdf)
Exercise sheet: Exercise 6 - El Farol Bar model (pdf). (note: this exercise will not be covered in detail in the lab)
NetLogo Models Library: Sample Models > Social Science > El Farol
Code: List example task 1 (netlogo), Simplified El Farol (netlogo), model with excerise 6 tasks completed (netlogo)
Paper: W. Brian Arthur (1994) “Inductive Reasoning and Bounded Rationality”, American Economic Review, 84,406–411. (pdf)

8. Evolution of cooperation

Evolution in ABM. Implicit and explicit fitness and selection. A simple evolutionary algorithm with explicit fitness, selection and mutation. Cooperation in the Prisoner's Dilemma (PD) game from game theory. Evolution applied to the PD game. Mean field game interactions model. Structured game interactions model on fixed lattice (spatial PD model). Evolving dynamic interaction structures using tags and a network. Repeated (or Iterated) PD games.

Slides: Evolution of Cooperation (pdf)
Code: mean field model (netlogo), fixed lattice model (netlogo), tag model (netlogo), network rewire model (netlogo), task 1 - modified mean-field model to implement games on a nework (netlogo), task 2 - modified lattice model to implement mutation and different payoffs (netlogo)
Paper: Nowak, M. & May, R. (1992) Evolutionary games and spatial chaos. Nature 359, 826-829 [lattice PD model]
Paper: Nowak, M. & May, R. (1993) The Spatial Dilemmas of Evolution. Int. J. of Bifurcation and Chaos, Vol. 3, No. 1. 35-78 (pdf) [further analysis with lattice PD model]
Paper: Szabo, G. & Fath, G. (2006) Evolutionary games on graphs (http://arxiv.org/abs/cond-mat/0607344) [an extensive survey paper.]
Book: Axelrod, R. (1984) The evolution of cooperation. Basic Books. [describes results of a tournament in which people submitted agents (programs) to play the iterated PD game]
Link: lecture slides (from another course) that describe Axelrod's iterated PD tournaments in detail (pdf)

9. Experimental methods and tools

Issues when designing ABM experiments. Verifcation, validation and replication of ABM. The BehaviorSpace tool in NetLogo. Vary a parameter in the Schelling model using BehaviorSpace. Load the results into a statistical package. Produce a visual plot of the effect of varying the parameter. How to run NetLogo experiments as a background headerless process on a server utilising multiple cores.

Slides: Experimental methods and tools (pdf)

Code: Segregation model with BehaviorSpace experiments discussed in the slides (netlogo). R scripts for creating 2D level plot from netlogo exp2 ouptut csv file (R script simple, R script fancy), Shell script for running headless experiments easily (nrl.sh).

Link: NetLogo BehaviorSpace tool manual pages: https://ccl.northwestern.edu/netlogo/docs/behaviorspace.html

Papers: