Cell Formation in a Batch Oriented Production System using a Local Search Heuristic with a Genetic Algorithm: An Application of Cellular Manufacturing System

摘要

Cellular manufacturing is a production strategy which is capable of solving certain problems in a batch manufacturing system. A batch manufacturing system produces some intermediate varieties of products with intermediate volumes. The volume of any single product may not be sufficient to justify the use of a dedicated set of equipments for its production. Under this condition, a few or several products will have to share the production resources to balance their utilization. Production equipment in batch manufacturing must be capable of performing a variety of tasks. One of the fundamental problems in cellular manufacturing system is the formation of part families and machine cells that is the cell formation. For cell formation the part families are identified that require similar processing on a set of machines. In turn, these machines are grouped into cells. Each cell is capable of satisfying all the requirements of the part family assigned to it. In this paper an approach is used to form the part families and machine cell in a batch oriented production system. This approach combines a local search heuristic with a genetic algorithm. The genetic algorithm is used to generate the sets of machine cells. The evolutionary process, embedded in the genetic algorithm, is responsible for improving the grouping quality of the sets of machine cell generated. When the machine cells are known, it is customary to assign a product to the cell where it visits the maximum number of machines. This is optimal to minimize inter cell movement. However, it does not guarantee good utilization of the machines within a cell. To overcome this problem, a local search heuristic, which takes into consideration both inter -cell movement and machine utilization, is applied. The heuristic consists of an improvement procedure that is repeatedly applied. The objective of the heuristic is to construct a set of machine/part groups and improve it, if possible. The heuristic feeds back to the genetic algorithm the grouping efficacy of the set of machine/part groups it constructs. It is continued until the optimum result is found. After applying this approach, the result with a grouping efficacy is higher than the existing initial machine part matrix. So this approach can useful in cell formation in any batch oriented production system.