Multi-Agent based Artificial War

摘要

MultiAgentSystem(MAS) is a discipline of Distributed Artificial Intelligence (DAI) where a number of agents collaborate for achieving some common goal. The land combat can be mathematically modeled as a non-linear dynamic system composed of many interacting autonomous and hierarchically organized agents adapting to a changing environment. In this paper we propose a simulative approach using the concepts of MAS for an artificial war. We have modeled the scenario by considering elements of non-linear dynamics, artificial life, complexity theory and MAS based simulation. The tool we adopted for our approach is NETLGOawell-knownmultiagent programmable modeling environment. One of the important factors of land combat is the attrition rate. We have presented a simulation model for analyzing the attrition rates using Lanchester law and Lotka-Volterra equations. We have validated the results obtained from our approach with the data available in the literature [3]. This simulative approach helps the department of defense for decision making in real time situations.