Hybrid Firefly Algorithms For Multi-objective Scheduling Problem

The essence of Multi-objective Permutation Flow Shop Scheduling(MPFSP)is to coordinate and optimize multiple function objectives under certain constraints.This kind of problem exists widely in practical intelligent manufacturing systems.It is of great significance to study this kind of problem for improving enterprise production efficiency and reducing enterprise production costs.MPFSP is a classical NP-hard combinatorial optimization problem.The study of MPFSP can provide effective suggestions for solving other combinatorial optimization problems.Aiming at the three optimization objectives of multi-objective permutation flow shop scheduling problem: minimizing total process time,minimizing delay time and minimizing maximum completion time,this paper proposes a hybrid Firefly Algorithms(HFA)algorithm based on Firefly algorithm(FA).The hybrid algorithm is based on Firefly algorithm and embedded in NEH model and probability model.The specific improvement works are as follows:Firstly,in order to ensure the diversity of the initial population and improve the quality of the initial population in the initialization process,this thesis adopts a combination of machine coding and NEH heuristic coding to initialize the initial population.Secondly,in order to accelerate the iteration speed of the algorithm,this paper records the information between the workpiece and the workpiece,and between the workpiece and the processing machine in the probability matrix.By combining blocks with information in probability matrix and using block mining to solve the problem,the convergence speed of the algorithm is improved and the dominant solution fragments in the feasible solution are retained.Thirdly,falling into local optimum is a problem faced by many heuristic algorithms.In this thesis,the global search strategy is used to increase the diversity of feasible solutions and improve the quality of solutions.In the repeated search strategy,K index is set up to record the update information of the best solution in the current maternal population.If K index exceeds the threshold,the search range is changed,the probability matrix information is cleared,the original accumulated information range is changed,and the information in different regions of the maternal population is searched,thus the global search ability is increased.In order to compare the performance of the algorithm,this thesis simulates Taillard case and Reeves case,and compares the number of non-dominated solutions(Number of Pareto solution NPS),Spacing metric SM(Spacing metric SM)and C index of each algorithm.Through a series of comparisons,HFA algorithm shows better performance in solving multi-objective permutation flow shop scheduling problems.