A modified imperialist competitive algorithm for scheduling single batch-processing machine with fuzzy due date

Abstract

This paper considers a fuzzy single batch-processing machine (SBPM) scheduling problem and aims to make it closer to a real-world application through a fuzzy set theory. For this purpose, jobs’ due dates are set to be fuzzy numbers and the membership function of a fuzzy due date assigned to each job represents the degree of satisfaction that a decision maker has with the completion time of that job. The objective is to maximize the total degree of satisfaction or equivalently to minimize the total degree of dissatisfaction over given jobs. Then, the fuzzy mathematical programming model is presented. To solve the model, we propose the imperialist competitive algorithm (ICA) and modify the assimilation policy (i.e., colonies moving) and imperialistic competition in order to overcome its immature convergence and improve its performances. Moreover, due to the significant role of parameters on the quality of random search algorithms, a robust calibration is applied on the parameters using the Taguchi optimization technique. To evaluate the proposed ICA, several random test problems are generated and its performance is compared to the traditional ICA, simulated annealing (SA), particle swarm optimization (PSO), genetic algorithm (GA), ant colony optimization (ACO), and earliest due date (EDD). The obtained computational results demonstrate the superiority and robustness of the modified ICA.