Discrete Differential Evolution Algorithm with the Fuzzy Machine Selection for Solving the Flexible Job Shop Scheduling Problem

Abstract

At the present, the government encourages entrepreneurs used the innovation and digital technology for enhanced the competitiveness and increased the productivity. To success with the mention above, the entrepreneurs need to adopt the intelligence of the computer in the business. As well as the manufacturing industry, the one important in the manufacturing is to schedule the job plan or the scheduling plan. Nowadays, the industry applied the computer to calculate the schedule and hold the machine balancing in the manufacturing process. The Flexible Job Shop Scheduling Problem (FJSP) is the complex problem which is found in the manufacturing processes. This problem occurs when the staff cannot maintain a balance between the jobs and the machines. In recent years, the researcher in the operation research areas attends to create the metaheuristic algorithm for solving the FJSP. The differential evolution (DE) algorithm is one of the computational algorithms which used to solve the operation research optimization problem. There are several research works which have been proposed; for instance. Mohamed et al. [1] applied the DE algorithm for solving unconstrained global optimization problems. In this algorithm, their proposed a new directed mutation rule based on the weighted difference vector between the best and the worse solution. The local search is utilized to enhance the search capability and to increase the convergence rate. Furthermore, a dynamic nonlinear increased crossover probability scheme is proposed balance between the diversity and the convergence rate or between global exploration ability and local exploitation. The result of their algorithm indicates that the improved algorithm outperforms and is superior to other existing algorithms. Salehpour et al. [2] developed the new version of the DE algorithm with the fuzzy logic inference system. This paper uses a fuzzy logic inference system to dynamically tune the mutation factor of DE and improve its exploration and exploitation. A fuzzy system used to considering the variation, namely, number of generation and population. The results obtained show the really good behavior of the proposed method and comparison. Zou et al. [3] presented a Novel Modified Differential Evolution (NMDE) algorithm to solve constrained optimization problems. This algorithm modifies the scale factor of the original DE algorithm by an adaptive strategy. In the crossover operation of this paper, they use the uniform distribution when the stagnation happens to the solution. Moreover, a common penalty function method adopted to balance objective and constraint violations. Experimental results show that the NMDE algorithm has higher efficiency than the other methods in term of finding better feasible solutions of most constrained problems. Huang and Huang [4] proposed the DE algorithm with the ant system for solving the Optimal Reactive Power Dispatch (ORPD) problem. The purpose of ORPD is to reduce active power transmission losses and improve the voltage profile in the power systems. The step of this paper follows the original of DE algorithm. In the mutation process, this paper avoids falling into local minima and save more computational time by using the variable scaling mutation. They test the performance of the proposed algorithm on the IEEE 30-bus system. The experiment shown that, this paper obtains better results with lower active power transmission losses and faster convergence A RT I C L E I N F O