Robust Intelligent Construction Procedure for Job-Shop Scheduling

This paper proposes a robust intelligent technique to produce the initial population close to the optimal solution for the job-shop scheduling problem (JSSP). The proposed technique is designed by a new heuristic based on an intelligent skip from the primal point of the solution space to a better one that considers a new classification of jobs on machines. This new classification is named mPlates-Jobs. The main advantages of the proposed technique are its capability to produce any size of the initial population, its proximity to the optimal solution, and its capability to observe the best-known solution in the generated initial population for benchmark datasets. The comparison of the experimental results with those of Kuczapski's, Yahyaoui's, Moghaddam and Giffler's, and Thompson's initialization techniques, which are considered the four state-of-the-art initialization techniques, proves the abovementioned advantages. In this study, the proposed intelligent initialization technique can be considered a fast and intelligent heuristic algorithm to solve the JSSP based on the quality of its results.