Research On Hybrid Flow Shop Scheduling Problem By Improved Whale Swarm Algorithm

Production scheduling is one of the key factors affecting the production efficiency and production stability of manufacturing enterprises.The Hybrid Flow-shop Schedualing Problem(HFSP)is an extension of the flow-shop scheduling problem.It exists widely in the fields of machinery,chemical engineering,logistics and medical.It belongs to the NP-Hard problem and has important theoretical research and engineering applications.In most practical productions,in addition to conventional workpiece and machine constraints,some special resource constraints need to be considered,such as limited buffer.This thesis studies HFSP and HFSP with limited buffers(LBHFSP),and uses the Improved Whale Swarm Algorithm(IWSA)to solve HFSP,LBHFSP,and multi-object LBHFSP.An IWSA was designed to solve the HFSP.Firstly,a hybrid initialization method is used to improve the quality of the initial solution.Then,based on the characteristics of HFSP,a new encoding-decoding method is designed,and the distance calculation method between individuals is redefined.Second,the individual movement rules are improved.The neighborhood search strategy is introduced to improve the algorithm search ability.Finally,experimental results validate the effictiveness of the proposed algorithm.As for IWSA for single-object LBHFSP,firstly,the mathematical model of LBHFSP is established.Then,the greedy lag decoding,neighborhood search,and inter-generation deduplication strategies suitable for solving LBHFSP are designed.Finally,experimental results validate the effictiveness of the proposed algorithm.A multi-objective Whale Swarm Algorithm(MOWSA)was designed to solve the multi-objective LBHFSP.Among them,the introduction of new individual selection and elite retention strategies ensured the quality of the Pareto frontier finally sought.Finally,experimental results validate the effictiveness of the proposed algorithmFinally,the final chapter summarizes the thesis and proposes some future research directions.