From Conceptual Models to Simulation Models : Model Driven Development of Agent-Based Simulations

Every social issue surrounding us today stems from complex factors. Now we need to establish a trans-disciplinary approach so that we can examine the flow of interconnected phenomena. The solution lies with new methodology called agent-based modeling. This concept makes it possible to model our society as a system formed by an interaction between numbers of autonomous agents. In the current state in the study of simulating agent-based economic models, however, there is a problem that needs to be resolved. That is the absence of integrated environment to support a whole research process from conceptual modeling to simulation analysis. The problem did not become serious too much up to now, because the models were small-scale and for experimental use. It becomes, however, indispensable to resolve the existing problem, as the simulations come to be used practically in social science, policy analysis, and business field. In this paper, we propose a new development process, which we call “Model Driven Development”. To realize the process, we also propose tools that help us build the conceptual models and simulation models. Then, the simple trade model is introduced as the example in the last part of the paper. 2 Background