Research On Task Allocation Method For AGV Collaboration System Based On Heterogeneous MAS

In Multi-Agent System(MAS),the goal of task allocation is organizes the distributed intelligent agent,assign the task to the suitable agent through the interactive of agent,and then solve the problem which can not be completed by single agent.In the research of MAS task allocation,it's difficult to allocate the task to the most appropriate agent and take full advantage of the performance of the agent,because the efficiency of interaction and collaboration between the agents is so low due to the heterogeneity of the agents;on the other hand,in pervasive computing environment,agent capabilities will also change with the dynamic environment changing.How to allocate the task effectively become a serious problem.In order to solve the proble describe above,the research work in the thesis is as follows:Firstly,on the analysis of the feature of the agent capability,we proposed a model of agent behavior-executing capability.So we could describe the capability for the heterogeneous agent,and the model will be the foundation of task allocation.Secondly,on the analysis of the task allocation model based on the traditional contract net,a MAS task allocation model based on extended contract net was proposed.We improved the phase of the tendering and biding in the contract net based on the model of agent behavior-executing capability,and analysis the process of the MAS task allocation based on the extended contract net.At last,simulation system was developed for the AGV(Auto Guided Vehicle)in hospital medical waste transport system.In this simulation system,the model of agent behavior-executing capability was applied to the description of the hospital AGV,and applied the mechanism of MAS task allocation based on extended contract net to the problem of task allocation in the simulation system.Through the simulation system to verify the feasibility of the model and the effectiveness of the task allocation method.