Improving NSGA-? Algorithm And Applied To Flexible Job Shop Scheduling Problem(FJSP)

According to the actual production status and the characteristics of a flexible job shop in Dalian machinery factory,in full consideration of coordination and balance among multiple objectives,such as minimizing the maximum completion time,balancing the equipment use,minimizing total drag period and the total load of the machine,the maximum operating machinery load,production cost index and so on,a multiple objective flexible scheduling shop model is set up.The method based on working procedure and machine,and the insert type method is used for coding and decoding.Using the idea of uniform design in the prepossessing of the machine chain,so the uniform search of the multiple random direction can be ensured.At the same time,a double coding method is adopted to generate a certain amount of machine chain with the minimum operating time,so as to ensure the diversity of the population,and effectively prevent the population from premature convergence.A new idea of local crossover and mutation is proposed,which combining the processed machine chain with a randomly generated process chain to form an individual.In the process of genetic operation,all the machine chain are combined with the process chain.Only the chain of the process does crossover and mutation,to complete the search for the optimal solution.Using the optimal solution,the optimal solution and the convergence time of the initial machine chain to show the effectiveness of the improved method.Analyzing the crossover and mutation of genetic algorithm,to find out the advantages and disadvantages of each method,and improve the process of crossover and mutation aimed at its shortages.Because the elite retention strategy may lead to optimal solution convergence mistakenly converge to the first level non dominated drawbacks,the limiting function is used to guarantee the diversity of the selections,to avoid deviating from the real convergence surface of the optimal solution,resulting in partial convergence or premature convergence.The crossover and mutation only take place on the critical path,which can effectively reduce the maximum completion time.Solving the flexible production workshop scheduling problem of XX machinery factory with the improved NSGA-II algorithm in MATLAB.Then using the method of AHP to determine the trget layer of weights selected the optimal compromise solution after producing a set of Pareto solution.