The Research And Application Of Adaptive Differential Evolution-Genetic Algorithm In Flexible Job-Shop Scheduling

Since the 21st century,the market competition becomes increasingly fierce,in this marketing background,how to properly arrange the production plan,scheduling how to effectively scheduling is the key issue to be considered for enterprise to efficiently product.Whether job shop scheduling problems can efficiently obtain the optimal solution is the core production and operation,therefore,research on job shop scheduling problems,both in theory or reality aspects it has important significance.The application of efficient scheduling optimization methods and advanced technology,plays an important role to improve production efficiency,reduce production costs and other aspects for manufacturing enterprises.Flexible Job-shop Scheduling Problem(abbreviation for FJSP)is a lot of simplified model of practical production scheduling problem,it belongs to NP(Non-deterministic Polynomial)-hard problem in the calculation of the amount.But because it is closer to the actual situation of production,making it attracts domestic and overseas research focus in the scheduling fields.Currently,among many intelligent optimization algorithms for solving FJSP,the genetic algorithm(GA)has fast random search ability,robustness,simple process.strong scalability and other characteristics,and in solving nonlinear optimization problems,genetic algorithm showed its strong adaptability,global optimization and implicit parallelism,which it has irreplaceable advantages in the study of scheduling optimization,thus becoming one of the most widely used algorithm.Through consulting a large number of information and reading literature,knows that it has the quality of the initial solution of genetic algorithm on the low side,and in the late evolutionary algorithm convergence speed is slow,easy to fall into local optimum,resulting in low overall efficiency of the algorithm the two problems.Differential evolution(DE)algorithm as a kind of fast random search algorithm,in solving optimization problems reflect the strong local search ability,it can precisely cooperate with global search ability of genetic algorithm,complete optimization with high efficiency,improve the performance of the algorithm.So combining the advantages of the two algorithms,I proposed an improved adaptive differential evolution genetic algorithm.The improved algorithm firstly improved on the initial population,when choose the machine for individuals adopt the method of the roulette wheel selection strategy,to improve the quality of the initial population.Secondly,in the process of crossover and mutation,an improved adaptive crossover and mutation probability are proposed,make the algorithm convergence speed has increased significantly;Thirdly,introduced the local optimal decision index,determine whether the algorithm appeared the trend of premature convergence,after the judge into local optimum,implement differential evolution algorithm to jump out of local optimum,to explore the new solutions near the local extremum;Finally,through the simulation results,reflecting that the improved algorithm has strong advantages in the optimal value and convergence speed two aspects,simulate to achieve the job shop scheduling system based on improved algorithm,the improved algorithm is applied to the system scheduling,prove practicability of the algorithm.