Applying soft-computing techniques in solving dynamic multi-objective layout problems in cellular manufacturing system. Issue 1 (September 2016)

• Record Type: Journal Article
• Authors: Ghosh, Tamal; Doloi, B.; Dan, Pranab
• Is Part Of: International journal of advanced manufacturing technology. Volume 86:Issue 1/4(2016)
• Page Start: 237
• Page End: 257
• Publication Date: 2016-09
• Subjects: Multi-objective cellular layout -- Production uncertainty -- Material handling -- Proximity factor -- Inter-cell material flow -- Genetic algorithm -- Simulated annealing; Manufacturing processes -- Periodicals; Production engineering -- Periodicals; 670.427
• DOI: 10.1007/s00170-015-8070-6
• Languages: English
• ISSNs: 0268-3768

Abstract

Abstract The formation of machine cells to process subsequent part families is not the only goal of the designing of an efficient cellular manufacturing system (CMS). A competent layout of the newly acquired cells is also essential to restrict the total inter-cell material handling cost which is primarily significant with large production volume. Furthermore, in realistic industrial scenario, the uncertainty of product demand can influence the layout configuration to be altered from period to period. Albeit there are numerous articles exist in the domain of CMS research considering cell formation problems, layout issues have not been addressed significantly. Therefore, the aim of our paper is to portray a reformed mathematical model of the inter-cell layout design problem in dynamic production situation considering material handling cost and a modified proximity relationship of manufacturing cells. The proposed Quadratic Assignment Programming (QAP) model is combinatorial in nature and is difficult to solve using traditional exact solution methods. The state-of-the-art soft-computing techniques are extremely advantageous for such QAP paradigms. Thus, we developed an improved genetic algorithm (IGA) and a simulated annealing heuristic (SAH) to sort out the abovementioned problem. Due to the inadequacy of datasets, we formed small to large size test problems (6 × 6 × 2 to 24 × 24 × 10) in logical way to cater the purpose. The proposed algorithms are successfully employed to attain near-optimal solutions to the test problems. Computational results demonstrate the proficiency of the IGA over SAH for all the test problems in terms of solution quality and computational time. In addition, we conducted a statistical data analysis to validate the test results.