Agent-Based System-Dynamic Economic Modeling: a Pioneering Approach
Summary
This work offers a synthesis between the disciplines of agent-based modeling, system-dynamics and economics. A thorough analysis and critique of the prior efforts to synthesize economics and system-dynamics by the Post-Keynesian economist Steve Keen is presented. A C# .NET library based framework for the visualization, diagnostics and analysis of system-dynamic models is provided. The work further details a modeling method based on the separation between aggregates and individuals and between strategic, logical and ethical system-dynamic variables and demonstrates this method by presenting an evolutionarily-trained multi-agent model on its basis. This model is used to explore a hypothesis in the context of an alternative interpretation of Keen's Debt-Deflation Model.