An analysis and design framework for agent-based social simulation

Abstract

Agent-based modeling is one of the popular tools for analyzing complex social systems. To model such systems, social attributes such as culture, law and institutions need to implemented as part of the context of a MAS, independently of individual agents. In this paper, we present MAIA; a framework for modeling agent-based systems based on the Institutional Analysis and Development Framework (IAD). The IAD is a well established comprehensive framework which addresses many social attributes. To make this framework applicable to agent-based software implementation, we inspire from some of the detailed definitions in the OperA methodology. The framework covers the different types of structures affecting agents at the operational level; physical, collective and constitutional. Moreover, this framework includes the conceptualization and design of evaluation. An agent-based methodology has also been developed from the MAIA framework which consists of two layers. A conceptualization layer for analyzing and decomposing the system and a detailed design layer which leads to the implementation of social models. MAIA allows the balance of global institutional requirements with the autonomy of individual agents thus enabling system evolution and reflecting more of reality in artificial societies. © 2012 Springer-Verlag Berlin Heidelberg.