Research On Flexible Job Shop Scheduling Methods Based On Whale Swarm Algorithm

The multi-variety,low-volume production model make the production process become complicated and difficult to control.Efficient and robust production scheduling is the key to ensuring high efficiency and reliability of the production process and improving the core competitiveness of the enterprise.Flexible Job Shop Scheduling Problem(FJSP)is a scheduling problem extended from the job shop and is a typical NP-Hard problem.It has important engineering application and theoretical research value.However,FJSP only considers the limitation of machine resources.In most actual production processes,it is necessary to consider the constraints of various auxiliary resources besides machine resources.Therefore,the research results of the Resource Constrainted Flexible Job Shop Scheduling Problem(RCFJSP)are more instructive to the production process of the enterprise.Based on the analysis of domestic and foreign research status,this paper establishes the mathematical model of FJSP and RCFJSP,and the solution method of the single target FJSP,single target RCFJSP and multi-objective RCFJSP are studied respectively based on the Whale Swarm Algorithm(WSA).For solving the single-target FJSP,the WSA is improved.Firstly,the hybrid initialization strategy is adopted to obtain the high quality initial population.Then,the individual encoding method and decoding method are improved.Based on the encoding method,the distance calculation method and individual movement rules of the whale in WSA are improved.Finally,a neighborhood search strategy based on critical path information is designed to improve the local search ability of the algorithm.Through the test on international standard examples and comparison with other algorithms,the effectiveness of WSA in solving FJSP is verified.For solving the single-objective RCFJSP,The WSA is improved based on the feature of RCFJSP.The MinEnd2 initialization rule,the greedy forward decoding strategy and the critical path information acquisition method for RCFJSP are designed.Through Comparing experimental results with some state-of-the-art algorithms,the effectiveness of WSA are verified.For solving the multi-objective RCFJSP,the MOWSA is designed to solve the problem.Among them,Pareto external archive,fast non-dominated sorting,crowding distance and other mechanisms are introduced to ensure the quality of the final solution.Through the test on the international standard examples,it is verified that the performance of MOWSA is better than the comparison algorithm.Finally,the work of the whole dissertation is summarized and the future directions worthy of further research are prospected.