Agent-Based Modeling Course
Course Organization: This short course will be delivered by Dr Georgiy Bobashev, who is a senior statistician at RTI International.
Target Group: This short course is aimed at the audience of statisticians, analysts and researchers in social sciences who are interested in learning of the tools available for making estimates of risks, future outcomes and qualitative relationships.
Course Location: 219 Brinkhous-Bullitt at UNC
Course Time and Duration: The agent based modeling course will begin on Wednesday, March 28, at 9-11AM and last for 11 weeks, ending June 13, 2012.
Requirements: You are required to bring your own laptop or have access to a computer where you can download NETLOGO software for use in class.
Course Objectives: The short course will describe existing modelling methodology covering statistical models, Markov models, discrete event, system dynamics models, and agent-based models. We will discuss and illustrate the differences in modeling objectives and the applicability of each of the tools to achieve the objective. We will discuss approaches to model validation so that they are "trustable". The second part will focus on the basics of Agent-Based models and will involve model building and analysis.
Course Registration: NO PREREGISTRATION NECCESARY. ALL ARE WELCOME TO ATTEND.
Course Credit: TBN.
Masterclass course outline:
I. Introduction to modeling
1. What is modeling? Modeling objectives, and type of objectives: predict a number, make a decision, understand a relationship, estimate risk. Advantages and disadvantages of each model type
2. Model types: statistical, Markov, system dynamics, discrete event, and agent-based. Advantages and disadvantages of each model type
4. Examples of data driven and data-free models
II. Agent-based models (including lab examples)
5. ABM objectives and components (agents, rules, environments, networks). Develop your first ABM in NetLogo.
6. Standards and good modelling practices. Overview, Design concepts and Details (ODD) protocol Common technical issues to consider
7. Simulation scenarios and analysis of the results
8. Why should one trust your model? Uncertainty and validation. Interpretation of the results
III. Student feedback