| Shop scheduling involves the production planning,procurement,warehousing,sales and other operation management,as the core of the production system,job shop scheduling optimization can improve increase the manufacture efficiency and equipment utilization.Due to trend of individuation of production,and increasing diversified process route,it’s a pressing need to achieve small quantities for customized production effectively for industrial enterprise.Improving the flexibility of production system has become one of the main means to enhance the competitiveness of enterprises.Most of the research about shop scheduling focus on deterministic problems,which are assume that various parameters are specific for simulation.The fuzzy job shop scheduling problem in this paper can describe processing time and delivery time in manufacture with much accuracy,which cannot be accurately described within a certain range,will improve it.Research on job shop scheduling considered fuzzy and flexible separately had made great achievements,considering both of the two kinds of characteristics will make the problem becomes more complex.With scale and constraints increasing,the problem gets more complicated,mathematical programming,heuristic and other methods were restricted,hybrids among intelligent algorithms will contribute to solving the problem.The genetic algorithm has a number of advantages such as randomness,robustness,simple operation,etc.and has been used to combinate with other algorithms widely.A new algorithmp put forward in this paper was improved on the basis of combination of immune and genetic algorithms.In this paper,a multi-objective fuzzy flexible job shop scheduling model is proposed based on weighted target value for the flexible job shop scheduling problem with fuzzy processing time and fuzzy due dates.An improved immune genetic algorithm was put forward,and its’ implementation process to solve the problem is presented.The chromosome used real string as code,which is proposed by Mitsuo Gen.Extracting vaccine operation spontaneously with similarity inhibited.A new vaccination operation with simulated annealing was proposed,determine the allele of vaccine fragment through detection strategy before vaccination,the specific approach is to compare the new and old individuals of the corresponding gene bit,Strict control of replace if the probability is small,then decide whether should be vaccinate in accordance with the validity of it’s solution.If there is no change in several successive generations of vaccination,a further comparison to the best individual the other genes value in the gene position can make,determine whether it’s optimal gene or partial optimal.Vaccination operation by simulated annealing probability so as to rise above efficiently premature convergence and poor local search ability,adding memory storehouse to make up for the inflexibility of fixed cross and mutation.Finally,through solving the simulation example of the reference by proposed algorithm,the feasibility and effectiveness of this algorithm are verified.Then using the proposed algorithm to solve the Kacem target on Makespan and customer satisfaction,which often be regard as a standard example of fuzzy flexible job shop scheduling problem,the proposed algorithm improve remarkably ability for global search and convergence rate.Next,the processing time and machine load are taken as indexes to measure 8?8 and10?10 examples,and the results are compared with other algorithms in the literature,it works better or reasonably well at least.At last,the convergence of this algorithm is proved by Rastrigin function as Benchmark,which is extremely deceptive,and compared with other algorithms in the literature,the proposed algorithm was superior to others when solve the problem easily fall into local optimum.This optimization can avoid to stagnated at local convergence in earlier stage of procedure,meanwhile,make up for the defects of fluctuation occurred at late stage as it nears the optimal solution. |
PDF Full Text Request