Study on genetic algorithm (GA) approaches for solving Flow Shop Scheduling Problem (FSSP)

Abstract

The scheduling problem is known as one of the well-known optimization problems. It occurs in many situations of our daily-life applications, especially in industrial fields. One type of scheduling problem is called Flow Shop Scheduling Problem (FSSP). It belongs to the class of NP-complete problem. During the last decades, researches on exploring more accurate and efficient heuristic methods to solve hard optimization problems have taken considerable attention from researchers. Among them, GA has been one of the powerful and widely used algorithms. In this paper, we present two GA approaches to solve FSSP. The main objective is to investigate the effectiveness and the efficiency of GA based on different variations of the chromosome representation, referred to as the job-based GA (jb-GA) and machine-based GA (mb-GA). We conducted numerical experiments using standard test problems (Benchmark test problems). We also present the comparison of results with those given by another heuristic algorithm (NBH Algorithm) and the optimal solutions reported in the literature. Those demonstrate the jb-GA is more effective and efficient almost all of the time. The current limitation of this approach, like many other heuristic methods, is that it still sometimes gives the near-optimal solutions.