Parallel Machines Lot-Size Scheduling Problem Based On Biogeography-Grey Wolf Algorithm

With the maturity of industrial automation and Internet of Things technology,traditional manufacturing enterprises have gradually begun to carry out intelligent transformation,production efficiency as the core competitiveness of enterprises,the rational allocation and effective utilization of production resources are inseparable from it.At present,parallel machine production mode has a wide range of applications,such as injection workshop,light-emitting diode manufacturing,cigarette processing,etc.,parallel machine scheduling problem has become a key factor affecting production efficiency,the problem needs to meet the constraints of the actual working conditions of parallel machine shop,to find the best workpiece sequence and machine assignment,so that the relevant performance indicators are optimal,therefore,this thesis takes the parallel machine scheduling problem as the research object,and conducts singleobjective and multi-objective optimization research on the problem,and a scheduling system was designed by using the algorithms,and the main research content is as follows.To begin with,the thesis analyzed the research status of parallel machine scheduling problem and related algorithms,and a hybrid biogeography gray wolf algorithm is designed to minimize the makespan by taking the unrelated parallel machine lot-size scheduling problem as the object.The algorithm adopts a method of partition encoding by combining integer encoding in batch and machine partition and real number encoding in lot-size partition,and introduces the logistic strategy initialization in the lot-size partition,and iterates according to the wolf pack migration mechanism,adopts the species migration strategy in the batch and machine partition,adopts the adaptive catastrophe operator,and introduces the local reverse neighborhood search strategy.Then,considering the sequence-related machine switching time,an improved twostage multi-objective biogeographic algorithm is designed to minimize makespan and the number of switching between machines.Based on the single-objective algorithm,the fusion strategy by reversing column is used to generate the initial population and two-stage neighborhood search is introduced,the hypervolume evaluation is introduced in Pareto nondominated sorting,and then the machine-based matrix coding is used to optimize the processing sequence in the second stage of the algorithm to shorten the switching time.Finally,the effectiveness and superiority of the proposed algorithms are proved by the simulation experiments of benchmark case and cigarette processing case,taking the scheduling problem of foundry shop of an enterprise as an example,based on Python development platform,taking the Flask framework and SQLite database as the server,and the simulation optimization model of the proposed algorithms as the back-end program,a parallel machine lot-size scheduling system based on B/S architecture is developed.In summary,aiming at the problem of parallel machine lot-size scheduling,this thesis proposed a partition coding and sequence coding method based on machine matrix,species migration mechanism based on wolf pack,local reverse neighborhood search strategy,two-stage neighborhood search strategy and Pareto nondominated ranking based on hypervolume evaluation,respectively,single-objective and multiobjective algorithms are designed,and the effectiveness of the proposed algorithms were proved by the comparative experiment of benchmark case and cigarette processing case.Based on the above research results,the foundry shop scheduling system is developed,which makes the algorithm proposed in this thesis have practical engineering application value.