Development of capability-based virtual cellular manufacturing systems in dual-resource constrained settings over semi-distributed layouts

Virtual Cellular Manufacturing Systems (VCMSs) as a highly flexible manufacturing concept have been designed to improve the performance of classical Cellular Manufacturing Systems (CMSs) and job shop manufacturing environments by creating virtual grouping resources. The underlying concept as one of the main napplications of Group Technology (GT), which is based on batch production oriented, is particularly valuable during high demand variety and variability with increases in machines types required by parts for processing. This research develops a new system named Capability-Based VCMSs (CBVCMSs), which considers the overlapping machine capabilities, worker skills, and part process plans by defining Resource Elements (REs). Formation of CBVCMSs is performed through formulating a new Mixed-Integer Non-Linear Programming (MINLP) mathematical model in a static manner. A Goal Programming (GP) approach simultaneously considers several objectives and constraints while all components of the system are generated at the same time. Moreover, the formulated model is defined based on the design issues of Dual-Resource Constrained (DRC) settings, i.e., parts processing will be delayed if either workers, machines, or both are not available. Workers must be flexible since in DRC settings the number of machines exceeds the number of workers. In this research, workers are assumed to have different levels of cross-training (multi-level) and proficiency (heterogeneous). The performance of the developed CBVCMSs is improved by utilizing a novel layout namely Semi-Distributed Layouts (SDLs) using Genetic Algorithms (GAs). Since the cell formation problem in CMSs is NP-hard and this complexity will be increased for the development of CBVCMSs especially for DRC settings, Multi-Objective Tabu-Search (MOTS) is investigated to achieve global or near-to-global optimum solutions for the developed model. The results obtained from different test-problems are analyzed based on the objective function, traveled distance by parts and material flow, Cell Capacity Utilization (CCU), and System Capacity Utilization (SCU). Analysis of results illustrated the priority of CBVCMSs compared to the equivalent CMSs. Through one to one comparison of the CCUs, 8 out of 12 virtual cells belonging to four CBVCMSs were found to have a CCU in total 63.84 % higher than the equivalent classical cells. Considering the averages or the summations of the CCUs, all test problems (4 out of 4), have values in total 5.39 % greater than the classical CMSs. Moreover, by considering SCU as the comparative criterion, CBVCMSs outperformed the equivalent CMSs in all test-problems by 106.08% improvement in the whole of the system. Since the objective function of the model generated over SDLs is in average 18.42% smaller than the functional layouts, for all systems, SDLs give better arrangements in the formation of CBVCMSs compared with FLs. Moreover, comparison of the minimum material flows revealed that in 9 out of 10 comparisons,the developed systems generated over the SDLs gave in average 31.81% smaller values. In addition, by considering the developed system in DRC settings, in all test problems, the dissimilarity of parts assigned to a cell and load unbalances among cells decreased in average 12.41% and 41.48% respectively compared to the same system without DRC settings.

Preview PDF FK 2012 21R.pdf Download (1MB) | Preview

Actions (login required)

View Item