Research On Modeling And Solving Dynamic Job Shop Scheduling Problem

The competition of modern manufacturing industry is intense. The old modeof production is not suited for the requirements of modern production. Large-scalestudy of plant manufacturing problems becomes a current hot issue. However, thecomplex content and extensive scope of the manufacturing systems are still weakin research areas. The scheduling problem with the NP-hard has become abottleneck. Therefore, effective scheduling and optimization methods are the keytechnology and infrastructure of the modern plant manufacturing. The birth of theAgent technology offers a new way for shop scheduling research. The Agenttechnology is a hot topic in artificial intelligence research today. In this paper, theworkshop intelligent scheduling system combines the improved contract net basedon Multi-Agent and the cloud-adaptive genetic algorithm (CAGA). The mainresearch results of the paper are as follows:1. The traditional contract net protocol is improved. Although Contract NetProtocol is the main method of Multi-Agent, there are still many problems in thetraditional contract net. All these affect the actual collaboration process andefficiency completion of the task. In this paper, an improved Contract Net Protocol(CNP) with the global two-way， Multi-Agent System (MAS) based oncommunication model, which incorporated the local autonomy of workingmutually in consultation by negotiation. The efficiency of contract netcommunication is improved.2. In the third chapter, the multi-Agent system model is established whichbased on the shop floor scheduling and details analysis of the Agents functions isdesigned. The improved contract net protocol is achieved between the ResourcesAgent and the Equipment Agent, and the scheduling algorithm is designed basedon multi-Agent.3. In the fourth chapter, the manufacturing grid resources managementprocess and the resources scheduling goal is discussed. Based on the cloudadaptive genetic algorithm (CAGA), multi-objective and certain limits resourcesscheduling method is proposed for users. Then the convergence properties of theproof and the typical function are tested with cloud adaptive genetic algorithm. Atlast, the experimental results and performance comparisons show that the proposedmethod is both effective and efficient. 4. In the fifth chapter, the overall scheme of the shop scheduling system andthe function of each scheduling Agent are designed. The main interface ofscheduling system is devised on JADE platform. Part of shop scheduling, taskassignment and conclusion is obtained. Then complete the assignment distributionthrough resources Agent and equipment Agent, and with the process of thetender-bid. On the basis of the experiment results, the communication amount isdecreased and the cooperation and information flow time is reduced.