|Abstract or Summary
- Cellular Manufacturing Systems (CMS) have claimed many advantages over traditional job shop processes. Some of the advantages reported by several users of CMS are reduction in throughput time, reduction in WIP inventory, improvement in product quality, faster response time to customer orders, shorter move distances, increase in manufacturing flexibility, and greater job satisfaction. In its implementation, CMS organizes a production floor into manufacturing cells. Hence, the important issue that needs to be addressed first is the cell formation (CF) problem. CF deals with the identification of part families, machine groups, and allocation of part families and machine groups to cells or vice versa. In the past, most studies in CF have assumed that the location of a cell is known a priori and a unique route exists between two cells. However, in an actual manufacturing environment, alternative locations are available for locating each cell. Similarly, when the capacity of the material handler being used is limited, alternative routes may have to be used to move part loads between two cells.
In this research, the issues dealing with alternative cell locations and alternative routes of material handling equipment are investigated. In addition, several other important factors common to CF are also considered. These include machine capacity limitations, batches of part demands, non-consecutive operations of parts, and maximum number of machines assigned to a cell. A mathematical model is first formulated to represent the research problem. The model is a binary and general integer non-linear
programming model, and it belongs to the class of NP-hard problem. Therefore, a higher level heuristic algorithm, based on the concept known as tabu search, is developed to efficiently solve the problems with industry merit. Incorporating the features associated with the tabu search, resulted in developing six different versions of the heuristic solution algorithm. The six heuristics are tested on twenty small problems, and the quality of their solutions is evaluated by investing significant effort to find their optimal solutions. The evaluation shows that the heuristics are highly effective. The solutions obtained from the heuristics have average percentage deviation of less than 3% from the optimal solutions. The heuristics are also tested on their performances with medium and large problems.
By using a statistical experiment that is based on randomized block design, the performance of the six heuristics is compared. Three different problem structures, ranging from 4 parts to 30 parts and from 3 locations to 9 locations are used in the experimentation. The experiment reveals that in general, the tabu search based-heuristic using fixed tabu list size and long-term memory based on minimal frequency strategy is preferred to other heuristics as the problem size increases.