A decode-based chaotic adaptive differential evolution for fuzzy job-shop scheduling problem

As a scheduling problem, the job-shop scheduling problem has attracted much attention with practical significance. Due to the uncertainty aspects of human factors and machine failures, job-shop scheduling problems with fuzzy processing time (FJSPs) have been widely used in actual processing and production. However, exact methods can not provide acceptable solutions for large-scale FJSPs. With the development of evolutionary computation, many meta-heuristic algorithms have obtained successfully high-quality solution on FJSPs. Although meta-heuristic algorithms are able to generate acceptable approximate solutions, they are still limited by low convergence and problem constraints. In this study, a decode-based chaotic adaptive differential evolution (DCADE) is proposed to alleviate the limitation. It includes a chaotic search, adaptive parameters, and decoding strategy. The chaotic search is used to improve the convergence speed, and the decoding strategy aimed at FJSPs can improve the solution quality of DCADE on FJSPs. Extensive experiments are implemented to verify the performance of DCADE on eight FJSPs compared with five state-of-the-art algorithms. Besides, the ablation study and parameter analysis are executed to discuss the impact of decoding strategy and parameters. The comprehensive experimental results demonstrate the superiority of DCADE.