Research On Fuzzy Flexible Job Shop Scheduling And Optimization Algorithm

Flexible job shop scheduling problem(FJSP)exists in industry and manufacturing widely,and has become a research hotspot of scholars gradually.At present,most of researchers consider FJSP as a deterministic problem.However,due to the differences in skill levels of workers and the aging and wear of machines and tools,the processing time and other parameters are fuzzy.In addition,with the rapid development of collaborative production among enterprises,the research on distributed flexible job shop scheduling problem(DFJSP)can able to integrate resource and meet customer needs.Now the DFJSP lacks research on the uncertainty problem,which leads to a certain impact on the actual production process.Therefore,the main contents of this paper are as follows:(1)Two kinds of fuzzy scheduling problem models are established.Firstly,the processing time is described as triangular fuzzy numbers and the related algorithms of triangular fuzzy numbers are analyzed.Then,the fuzzy flexible job shop scheduling problem(FFJSP)model with the fuzzy processing time is researched and the makespan is taken as the scheduling objective.At last,the multi-objective fuzzy distributed flexible job shop scheduling problem(MFDFJSP)model with the optimization objectives of minimizing fuzzy makespan and total carbon dioxide emissions is established,and a turning off/on policy is adopted when the machine is during the idle time to reduce total carbon dioxide emissions.(2)An improved whale optimization algorithm(IWOA)for solving the FFJSP scheduling problem is designed.The mathematical model of whale optimization algorithm(WOA)is briefly summarized.To solve the FFJSP,a mechanism for mutual conversion between scheduling solutions and whale individuals is constructed.The global search,local search and random generation(GLR)initialization methods are used to improve the quality of the initial population,and the local search method is integrated into the WOA,which increases the spatial search ability of the algorithm.The best parameter combination of IWOA was selected by Taguchi orthogonal experiment method.Through comparative experiments,the effectiveness of the IWOA is verified,and the solution method of FFJSP scheduling problem is innovated.(3)A memetic algorithm for solving MFDFJSP scheduling problem is proposed.The framework of IMA is given.According to the problem characteristics and target characteristics,a fusion GLR initialization method and an initialization method based on the overall global minimum of carbon dioxide emissions are designed.In addition,the crossover operation,the mutation operation and the local search operation including four neighborhood structure are constructed.Fifteen benchmarks suitable for the MFDFJSP are designed,and the optimal parameter combination of the IMA optimization algorithm is selected according to the Taguchi orthogonal experiment.Through self-comparison,the advantages of the initialization operation and local search operation are verified.Compared with NSGA-Ⅱ and MOPSO,the effectiveness and superiority of IMA in solving MFDFJSP scheduling problem are proved.