A meta heuristic web based optimization tool for assembly line balancing problems

Abstract

Presently Assembly Line Balancing (ALB) problems are very common in many industrial systems and these problems are addressed based on an a set of assembly tasks assigned to an ordered sequence within the workstations. The purpose of this study is to investigate the use of heuristics and meta-heuristic in addressing Simple Assembly Line Balancing Problems (SALBP) and develop a webbased optimization tool based on heuristics and genetic algorithm (GA). This system was developed using Hypertext Preprocessor (PHP) and MySQL. The heuristic techniques used were longest operation time (LOT), largest candidate rule (LCR), and ranked positional weight (RPW). An improved fitness function based on the modified GA was proposed in this study as a means to avoid the problem of chromosome selection in classic GA and to find a faster ALB solution in an internetenabled environment. The effect of improved fitness function and classic fitness function of modified GA on the performance of the developed web-based system was studied and the effectiveness and inadequacies of modified GA are presented. Comparison of the techniques will be determined and analysed based on the effectiveness of each techniques. The result of the standardised datasets indicated that the performance of the modified GA was superior compared to the other heuristic techniques based on the ALB results. In addition, the limitation of the web computation time for web-based optimization tool was also investigated. The results demonstrated that in most cases, the modified GA is able to produce ALB solution that can work within the limitation of the computation time. Furthermore, the system has been developed to benefit the industry by assigning a set of assembly tasks to workstations according to their main constraints as well as reducing the number of workstations needed.