Multi-Agent Based Preemptive Job Shop Scheduling

The manufacturing enterprises want to gain advantage in the market,they muststrengthen the construction of information technology, and ERP is a measure of anenterprise’s information degree. In the modules of the ERP systems, productionscheduling system is one of the most critical modules,it occupies a core position in theERP system. Establish sound and efficient production scheduling system is the key tomanufacturing enterprises to achieve stable, safe and effective operation. The job shopscheduling problem (JSP) is one of the most typical and most important schedulingproblem,it has been widely studied and we have obtained a wealth of theoretical resultsin the JSP. The preemptive job shop scheduling problem is an extension of the classicaljob shop scheduling problem, change the constraints of a process of the workpieceduring the machining process can not be preempted, so PJSP problem is more complexthan the JSP problem and JSP problem has proved to be an NP-hard problem. Bystudying the literature,we find that due to the complexity of the preemptive schedulingproblem, the studying about the preemptive scheduling is still relatively small, thedynamic preemptive job shop scheduling problem more less. And, as far as author know,the preemptive job shop scheduling problem is mainly some optimization andapproximation/heuristic methods, but these methods have some obvious shortcomings,are not adapted to PJSP problem. In the literature, there is no application of the agentapproach to solving the preemptive job shop scheduling problem. Multi-agent system asan important method in artificial intelligence, it can intelligent control andimplementation of operational disruptions, and therefore multi-agent technology is anatural solution strategy to solve PJSP problem in practical workshop, by theintroduction of multi-agent technology into PJSP problem, it can prove that themulti-agent approach can validity address such a complex scheduling problem.In this paper,the multi-agent system uses weights negotiation strategy.By studyingthe preemptive job shop scheduling system and multi-agent system technology,authorbuild based multi-agent preemptive job shop scheduling system model, and by theArena simulation software,build a based multi-agent preemptive job shop schedulingsystem simulation model,also build a based dispatching preemptive job shopscheduling system simulation model, last,by simulation experiments,it prove basedmulti-agent preemptive job shop scheduling system has better performance thandispatch rules with the objective function of the standardized lag (Normalized Mean Tardiness, NMT),Meanwhile,author do Some improvements about the basedmulti-agent preemptive job shop scheduling system, make it more close to the actualproduction scheduling system. In the last part of the article, in order to enable the designof multi-agent c preemptive job shop scheduling system to better reflect the real-timeshop floor status changes, to update the evaluation index of each agent and thecomprehensive evaluation index, author uses t the strategy of the timer interrupt.