The Application Of Improved Genetic Algorithm In Multi-objective Job Shop Scheduling

Job shop scheduling system for the past, only to adapt to a specific shop environment and a single objective of the evaluation criteria of the scheduling options, such as only getting the shortest time, lowest cost, load balancing equipment etc. Designed and realized a common multi-objective job shop scheduling system. According to a variety of real-time business information, such as real-time inventory information, in-progress information, plant capacity information, dynamic comprehensive control, generate real-time high-shop scheduling programs. At the same time, you can choose cost according to the actual control of the production process to make the final scheduling programs to meet the general assessment criteria, and taking into account the various objectives to achieve as much as possible on better conditions. This system, besides considering the various constraints in parallel, simultaneously improves suitable occasion of the system. The existing research work or study is mostly the scheduling process, or assembly of the scheduling research, it is difficult for production patterns are mixed types of manufacturing industry. To address the above issues, required not only a good solution to multi-objective optimization algorithm, but also an appropriate solution for a practical workshop.In this paper, a good parallelism of genetic algorithms can optimize the characteristics of multi-objective problem proposed, multi-objective genetic algorithm based on random weight preference is proposed, which solve multi-objective genetic algorithm that is difficult to make a choice. Algorithm uses a random weight ease of use, as well as multi-search features, and preferences of the leading light of the evolution of these characteristics to filter the solution to be Pareto. Algorithm overcomes the blindness of the random weighting method in search , but also solves a result of totally dependent on preference information brought in the calculation of the complexity and improve the multi-objective genetic algorithm to generate Pareto solution performance.For the multi-objective job shop scheduling actual situation in mixed production patterns, this article has designed a new application for mixed production patterns, and present a optimization model that can simultaneously concerned about multi-target. Compared with traditional multi-objective optimization algorithm, the algorithm has good versatility and optimize needed multi-target, and therefore the model has better practicability.This paper introduces application platform key technologies that develop the pattern of mixed-production job shop scheduling and applied the new algorithm to scheduling platform. In final, the final result is feasible and effective by the practical problem solving.