Invisible, Instantaneous and Initialised: Case Studies of the Value of Odd Assumptions in Agent Based Models

We are all used to models that take place on the "FlatWorld" of NetLogo. Despite variations in the use of networks, resources, artefacts and agent architectures across models, the "literal" interpretation of FlatWorld can create problems in reconciling ABM with real social situations and data. The bulk of the paper involves the discussion of two case studies (initialising an ABM involving matching of workers to jobs and modelling the evolution of spatially located social networks) where making apparently odd assumptions (turning agents invisible and running the simulation without "time") actually helps to reconcile ABM more effectively with social science and its data.