Quality Inspection Scheduling Problem with Adaptive Hybrid Genetic Algorithm

Abstract

In order to improve the efficiency and accuracy of quality testing of electronic meters, and replace the existing manual scheduling mode, automatic quality inspection job scheduling has become a natural choice for laboratories. However, different from the existing flexible job shop scheduling problem (FJSP), the quality inspection scheduling problem (QISP) has obvious differences in the correspondence between inspection tasks and batches of samples, solution constraints and the problem scale, making the existing scheduling algorithm unable to be directly applied. This paper proposes a new mathematical model for the quality inspection scheduling problem, and an adaptive hybrid genetic algorithm (AHGA). During the decoding operation, several neighborhood search strategies and heuristic rules are presented to ensure the feasibility of the solution. The elite retention strategy is introduced in the selection operation to relieve the loss of high-quality solutions. In terms of genetic operators, a mechanism for adaptive adjustment of operator crossover and mutation probability is designed to balance the global search and local search capabilities. The simulated annealing mechanism is used to speed up the algorithm's convergence and ensure the diversity of the population. Finally, the feasibility of the model and the algorithm is verified on the dataset of a state grid company.