A Python-Based MPI Framework for Exploring an Adaptive Fuzzy-Agent Approach to Simulating Large-Scale Non-Cooperative Games

In this article, we describe how to construct a large scale simulation system using the standard message passing interface (MPI) framework which can effectively explore the simulated players' strategy search spaces (i.e., to identify "good" strategies within particular "games" out of large sets of potential strategies) using genetic algorithms. We demonstrate how to create "intelligent" players who are capable of adapting their behaviors as the game evolves, given the problematic nature of identifying "good" strategies a priori using fuzzy logic. We prove these two concepts by building a scalable predator and prey simulation framework.