A Biologically-Inspired Platform for the Evolution of Communication in Multi-Agent Systems

Abstract

This research effort seeks to address the problem of limited communication in multi-agent systems by the use biologically-inspired implementations of intelligent control systems. A detailed simulation platform is built and tested. The platform contains many user-configurable parameters and is intended to be used as a general purpose research tool for the study of the evolution of communication in multi-agent systems. The user-configurable parameters include those that modify the agent environment, the structure of agent communication abilities, and various simulated biological selection forces and influences including natural selection, cultural transmission, and sexual selection. The simulation involves a predator-prey (pursuer-evader) environment in which evolving predators seek to capture prey and thus increase their fitness. Agents are given the ability to communicate, but are not forced to do so. Simulation results are presented to demonstrate the usefulness and abilities of this tool.