Research On Job-Shop Scheduling Under Uncertainty Based On Genetic Simulated Annealing Algorithm

With the development of market diversification and personalized needs, manufacturing mode tends to be more and more flexible. In order to improve the competitive ability of the enterprise, job shop scheduling as the important part of production management system has been heated up in research. Ignoring uncertainties and disturbance factors of real production process, the current research mostly is based on static assumption. Therefore, it is significant to research job shop scheduling problem under uncertainty both in theory and reality.Based on totally consideration for uncertainties and disturbances existing in real job shops, this paper established its uncertain job shop scheduling model and finally got the optimization by the use of improved genetic simulated annealing algorithm. The main research and results can be summarized by the following points:(1) Considering the complexity of real job shop, firstly a detailed analysis was carried out on uncertainties in the production process and methods to describe it and establish uncertain scheduling model. Through analysis and comparison, regarded uncertain processing times and duration as fuzzy parameters in scheduling. And then studied on the problem how to deal with uncertain disturbances of job shop based on dynamic scheduling strategies.(2) Designed a hybrid genetic simulated annealing algorithm with a combination of genetic algorithm and simulated annealing algorithm features. According with the demands of complex job shop, a further research to improve on the points such as coding, decoding, crossover, mutation and optimization strategy made the hybrid algorithm more advantageous.(3) Established a multi-objective job shop scheduling model with uncertain parameter, using the hybrid simulated annealing algorithm into the model, and obtained an initial scheduling scheme. Then on the basis, uncertain disturbances such as emergency order insert, machine failure, order cancellation and duration changes was involved in uncertain model, and through the comparison of example, the results verified the feasibility and effectivness of the model.(4) By research CSB real workshop of BSPT company, established uncertain dynamic scheduling model. And the results showed that the model and algorithm were feasibility and can be well used into real workshop.