A Hybrid PSO Based Algorithm for Solving the Machine-Part Cell Formation Problem

Abstract

Cellular Manufacturing (CM), an approach primarily based on the concept of Group Technology (GT), is one of the recent trends that help the manufacturing industry in reducing manufacturing cost and increasing productivity while maintaining quality. The idea of manufacturing parts in dedicated cells is beneficial as it results in increased manufacturing quality and reduced lead times. However, implementation of such a system in a real-life situation is always a challenging task. To overcome this challenge, several techniques, including AI-based approaches, have been developed over the years and regularly reported in literature. A very small portion of these approaches are utilizing Particle Swarm Optimization (PSO) in standard or hybrid form, whereas a larger chunk is either GA-based or utilizing other heuristics. To test the effectiveness of PSO while handling the Machine-Part Cell Formation (MPCF) problem in a CM environment, initially a standard PSO is developed during this research. Later, the same is hybridized with a Local Search Heuristic (LSH). The results of both standard and hybrid PSOs, developed during this research, are compared with the corresponding GA based methodologies, already available in literature. Computational results show that the GA based approaches have been outperformed both in terms of accuracy and computational effort. Further comparison of the results generated by the Hybrid PSO (HPSO) with several other techniques also shows that HPSO is either more or, in few cases, equally effective.