The Research On Job Shop Scheduling Based On An Intelligent Optimization Algorithm

In modern manufacture, shop scheduling is most important to every aspect of theproduction and management such as promoting the quality of production, decreasingcost, promoting the rate of production, responding to demand at the first time and soon. Many people has studied on Shop scheduling for many years, but there is muchwe need to do.In this thesis, the classical Job Shop, a kind of flexible Job Shop and the Job Shopunder uncertainty are studied through different optimal algorithms.In the first chapter, the description and classification of production schedulingand shop scheduling is promoted, and then almost all the optimal algorithms appliedto shop scheduling are summarized.In the second chapter, the classical Job Shop respectively based on adativesimulated annealing and ant algorithm is studied. In the part of adative simulatedannealing, a special updating function of temperature is promoted, when the searchfalls in a local optimization, the temperature will be increased after a sort, thereforethe search can leave the local optimization to some extent. Temper is also applied tothe adative simulated annealing. In the part of ant algorithm, the basic principle,model and the evolution of ant algorithm are introduced firstly, then a special antalgorithm applied to the classical Job Shop is promoted. In the ant algorithm, theearliest allowed processing time(EAPT) is considered as the heuristic information, allant start at different place and can select the latter operation decidedly or stochasticlyaccording to situation. After that, the process of the ant algorithm is promoted. At last,several kinds of benchmarks based on this two algorithms are studied and this twoalgorithms are compared.In the third chapter, a kind of flexible Job Shop is studied. At first, the model ofthe flexible Job Shop is introduced, then the principle, structure, feature and processof the genetic algorithm is introduced. After that, three genetic algorithms are appliedto three sub-problems of the flexible Job Shop, and the process is promoted. Lastly,two examples based on the algorithm is studied.In the fourth chapter, a kind of stochastic Job Shop in which the processing timesof all jobs are uniformly distributed variables. Firstly, the model of the stochastic JobShop is introdued, then the thought of the algorithm is promoted, it concludes threesteps: obtaining the datum needed in the neural network through stochastic simulation,training the neural network through genetic algorithm. after training the neuralnetwork, inserting the neural network into another genetic algorithm and applying thesecond genetic algorithm to Job Shop. Secondly,the principles of stochastic simulationand neural network are introduced, then the algorithm applied to the stochastic JobShop and its process is promoted. Lastly, a exmple is studied.In the fifth chapter, all the work in this paper is summarized and the prospectabout Job Shop is pomoted.