Steps to build an agent-based model
- Express the problem
- Search for theoretical issues or past researches on it
- Define your model
- Validate the model
- Run the model and organize results
- Discuss on the on the results
Gilbert and Troitzsch. 2006
As you can see these steps are pretty like any other types of social research. But main differences may rely in details of each step. Then I will try to provide some notes on each step.
Express the problem
The first question to ask here is: what kind of social problems can be addressed using agent-based modeling? The is not simple, but looking at main research stream in this field we can name few keywords which are very frequent: evolution of norms, cooperation, social networks, game theory, emergent phenomena, trust, micro-macro relationship, and so on.
If you are serious to enter this kind of social research, then check at least titles of most cited in this field. You can find then in Google Scholar. Check abstracts of some papers to understand here how we see the society.
Like other fields of study, agent-based modelers in sociology are a community/ communities. Searching on the net you can find mailing lists, useful websites, free book chapters and papers and many friendly scholars who may welcome your questions. But keep in mind: don’t pay your asking coupons for those elementary questions. Study, think and then ask!
To be continued …
