Research On Job Shop Scheduling Design And Simulation Based On Improved Particle Swarm Algorithms

Job shop production line is an important means for enterprise to improve the quick reactionability, satisfy the diversification demand of products, reduce production costs, and improveenterprise benefit. The manufacturing cycle of production line and machine availability of productionline have a close relationship and interaction, which have an essential effect to the effective utilizationof job shop production line. However, most studies only consider the manufacturing cycle, ignoringthe machine utilization. This article carried out studies based on improved particle swarmoptimization for job shop sequencing and simulation. The major research work done and resultsachieved are as follows:(1) Systematically test the effect of population size, inertia weight, accelerated factor1and factor2the four control parameters to the performance of particle swarm algorithm, and obtain the effect ofthe change of control parameters to the particle swarm algorithm(2) Proposed three improved particle swarm optimization: new adaptive PSO which the inertiaweight and two constants both have a adaptive change as the change of the particle fitness value;genetic particle swarm optimization which the population is divided into several sub-populations,populations of the first pair genetic algorithm, and then through the genetic operation as the initialpopulation of sub-populations, and finally PSO operation; Cooperative PSO which reference tocollaboration technology, niche technology, genetic algorithm crossover and mutation techniquesbased on niche technology. Test and compare the performance of the algorithms including basicparticles swarm algorithm, adaptive particle swarm optimization, genetic algorithm and particleswarm algorithm based on niche technology, obtain that both the convergence reliability andconverging velocity of adaptive particle swarm optimization, genetic algorithms and particle swarmbased on niche technology are all better than basic particles swarm algorithm(3) Establish a job shop scheduling optimization model according to the optimization model, dodetailed design of the three improved particle swarm algorithms including the adaptive particle swarmoptimization, genetic algorithm and particle swarm based on niche technology, and apply them to thisoptimization model, compare their algorithm performance in this specific application.(4) Develop a set of Job Shop Scheduling Optimization System by using of Matlab, which cantest the effect of control parameters to the performance of particle swarm algorithm, and can also testthe performance of the algorithms including basic particle swarm optimization, adaptive particle swarm optimization, genetic algorithm and particle swarm based on niche technology, and apply theabove algorithms to the optimization of job shop scheduling problems.(5) Establish a stochastic model of job shop scheduling production line by the simulationsoftware Arena, add various uncertain factors which can meet in actual production, such as uncertainproduction time, uncertain transport time between machines, uncertain machine breakdowns and thesubstandard product of work piece. The simulation result shows that the load of the machines on theproduction lines is still very uniform, but the manufacturing cycle prolonged.