Intelligent optimisation for multi-objectives flexible manufacturing cells formation

Abstract

The primary objective of conventional manufacturing cell formation typically uses grouping efficiency and efficacy measurement to reduce voids and exceptional parts. This objective frequently leads to extreme solutions, such as the persistently significant workload disparity among the manu­facturing cells. It will have a detrimental psychological impact on operators who work in each formed manufacturing cell. The complexity of the problem increases when there is a requirement to finish all parts before the midday break, at which point the formed manufacturing cells can proceed with the following production batch after the break. This research examines the formation of manufacturing cells using two widely recognized intelligent optimization techniques: genetic algorithm (G.A.) and particle swarm optimisation (PSO). The discussed manufacturing system has flexible machines, allowing each part to have multiple production routing options. The optimisation process involved addressing four simultaneous objectives: enhancing the efficiency and efficacy of the manufacturing cells, minimizing the deviation of manufacturing cells working time with the allocated working hours, which is prior to the midday break, and ensuring a balanced workload for the formed manufacturing cells. The optimisation results demonstrate that the G.A. outperforms the PSO method and is capable of providing manufacturing cell formation solutions with an efficiency level of 0.86, efficacy level as high as 0.64, achieving a minimum lateness of only 24 minutes from the completion target before midday break and a maximum difference in workload as low as 49 minutes.

Keywords

• flexible manufacturing cells formation
• genetic algorithms
• multi-objectives intelligent optimisation
• particle swarm optimisation