Efficient combined immune-decomposition algorithm for optimal buffer allocation in production lines for throughput and profit maximization

• Published in: Computers & operations research Vol. 37; no. 4; pp. 611 - 620
• Main Authors: Massim, Y., Yalaoui, F., Amodeo, L., Chatelet, E., Zeblah, A.
• Format: Journal Article
• Language: English
• Published: Kidlington Elsevier Ltd 01.04.2010; Elsevier; Pergamon Press Inc
• Subjects: Algorithms; Applied sciences; Artificial immune system; Assembly lines; Average throughput; Buffer inventories; Buffers; Computer Science; Decomposition; Decomposition method; Exact sciences and technology; Inventory control, production control. Distribution; Mathematical programming; Operational research and scientific management; Operational research. Management science; Operations Research; Optimization; Optimization algorithms; Profit maximization; Studies; Work in process; Work inventory; Work inventory; Average throughput; Buffers; Optimization; Artificial immune system; Decomposition method; Costs; Optimal algorithm; Performance characteristic; Production line; In-process inventory; Buffer system; Economic sciences; Immune algorithm; Production management; Optimal allocation; Profit
• ISSN: 0305-0548; 1873-765X; 0305-0548
• DOI: 10.1016/j.cor.2009.06.016

Adequate allocation of buffers in transfer lines is crucial to the optimization of line throughput and work in process (WIP) inventory. Their optimal allocation is subject to specific constraints, associated costs, and revenue projections. In this paper, we implement a combined artificial immune system optimization algorithm in conjunction with a decomposition method to optimally allocate buffers in transfer lines. The aim of the buffer allocation problem (BAP) is to achieve optimal system performance under buffers space constraints. Maximizing line throughput does not necessarily achieve maximum profit. In this study the immune decomposition algorithm (IDA) is used to determine optimal buffer allocation for maximum line throughput and maximum line economic profit. Results of extensive series of tests carried out to compare, in production lines with different characteristics, the performances of the proposed method and those of other algorithms are presented.