ABM workshop: Agent-based modelling for theory development

ISEM 2019, Salzburg, Austria
Facilitators: Volker Grimm, Uta Berger

Agent-based modeling is a computational approach to analyzing the behavioral decision of adaptive agents (plants, animals, humans, institutions) and investigate the behavioral and system-level properties that emerge. The approach is widely applied in many disciplines, including ecology, geography, and economics, for both particular case studies and strategically as virtual laboratories for developing theory at agent and system levels. ABMs are thus essential tools that mechanistically complement the conceptual framework of Complex Adaptive Systems.

Date: October 1, 2019

Venue:

10:00-11:30

Rationale and methodology of ABM

11:45-13:00

NETLOGO part 1 – the most common platform for developing ABM

13:00-14:00

Lunch

14:00-15:15

NETLOGO part 2 – how to do doing simulation experiments

15:30-16:45

Model tests and applications

Layout of course:

  1. Rationale and methodology of ABM
    1. Rationale and benefits of agent-based models
      • What ABMs are and when and why they are used
      • What agent-based modeling is, and how it works
    2. Methodology
      • What decisions you have to make to formulate an ABM
      • The role of observed patterns for designing and parameterizing ABM
    3. NETLOGO part 1 – a programming platform to implement ABM
      • Key elements and concepts
      • Getting started: implement your own model
  2. NETLOGO part 2 – simulation experiments with NETLOGO
    1. BehaviourSpace – to set up simulation experiments
    2. BehaviourSearch – using genetic algorithms for parameter optimization
  3. Model tests and applications
    1. From sensitivity analysis to robustness analysis
    2. Boosting your analysis through R-NETLOGO-Connection (nlxr package)

The course consists of short lectures and mainly exercises using your own Laptop and the software NetLogo. No prior knowledge on programming or modelling is required – just be open and curious! You will need to – prior to the course – download and install NetLogo: https://ccl.northwestern.edu/netlogo/6.0.4/.

Spending 1-2 hours going through the NetLogo tutorials would be useful.