Research And Implementation Of Multi-Agent System Framework Model With Flexibility For Decision Simulation

In the information age, the decision behavior presents the characteristics of variousrules, various cooperative processes and various application fields. Agent baseddecision simulation is an important way to study decision behavior，improve decisionprocess and raise decision standards. How to build up a MAS(Multi-Agent-System)framework of decision simulation to achieve effective simulation of dynamic and varieddecision behavior is a vital research question currently.This paper introduces agent techniques into decision simulation field, studies andpresents a MAS framework model with flexibility mechanisms. Firstly, after theanalysis of existing MAS framework and corresponding technology, the paper gives theframework model’s design principle and implementing method from static structuredesign, dynamic working mechanism and the method of interactive cooperation.Secondly, based on script technology，metadata technology and framework model’soperation mechanism, the paper expounds the flexibility mechanisms of the frameworkmodel, including the strategy of extended application of the framework, the method tomodify decision rules dynamically and control mechanism to change cooperativerelationship smoothly. Lastly, the key modules and core mechanism of the frameworkare realized as development tool and simulation platform.The framework model is applied in the decision simulation of K field and D fieldseparately. The research and analysis of the experiment procedure and results provesthat the framework model is feasible and effective, at the same time it verifies theflexibility mechanisms of the framework model.