On Multi-Agent Based Simulation

Agent-Based Computational Economics (ACE) and Agent Based Simulation (ABS) are two relatively young fields of research that lie on the boundary of social science, computer science and the cognitive sciences (AI and cognitive psychology). Originally these fields find their origin in computational physics, where multi-agent systems (MAS) are used to study models with heterogeneous agent populations. As AI researchers were looking for new applications for their artificial agent models, these could be readily found within the confines of the social sciences, in particular in economics and game theory, that are already strongly founded in analytical and computational methods. For example John von Neumann was a strong proponent of the idea of replacing Homo Oeconomicus with so called Homo Algoritmicus, focussing more on the process of rationality rather than on the outcome. The proponents of agent-based simulation are therefore sometimes referred to as Simulationists since the research is mainly centered around simulation studies, instead of mathematical analysis. At the boundaries where the orthodox deductive-analytical research program and the simulation approach touch on issues that have usually solely been studied within the orthodox research program, the first nowadays finds more competition from the second during formal and informal meetings than perhaps a decade ago. Although this may cause heated debates among researchers, it may turn out that there is much to learn on both sides of the divide as the boundaries between the two approaches begin to fade and a new synthesis is formed between the more rigorous mathematical analysis of orthodox neoclassical economics and the more