Flexible job-shop scheduling based on improved NSGA-II algorithm

Abstract

Flexible job-shop scheduling (FJSP) has the features of complex constraints, contradicts or conflicts between different objectives and unsatisfactory solution results. According to the actual production of the workshop, an improved NSGA-II algorithm is proposed to solve multi-objective FJSP problem, multiple objectives such as makespan, tardiness time, bottleneck machine load, total machine load and production cost are included in the algorithm. On the basis of NSGA-II algorithm, an extended MSOS operation coding is used to encode the operation substring and machine substring, the binary tournament and the optimal 1% elite strategy are integrated to select individuals with high fitness value, POX crossover and two-point crossover are used for cross operation, reverse mutation and intelligent mutation are used for mutation operation. The improved NSGA-II algorithm is used to solve the benchmark problems, the result is better than compared algorithms. Finally, take water turbine manufacturing process as an example, the improved NSGA-II algorithm is used to solve FJSP problem in machining workshop, a large number of satisfactory Pareto solutions are obtained. TOPSIS is used to evaluate the Pareto solutions, and the most satisfactory solution is chosen for decision makers.