Article ID: | iaor1992143 |
Country: | United States |
Volume: | 37 |
Issue: | 5 |
Start Page Number: | 587 |
End Page Number: | 606 |
Publication Date: | May 1991 |
Journal: | Management Science |
Authors: | Matsuo Hirofumi, Sullivan Robert S., Shang Jen S. |
Keywords: | production, programming: linear |
This paper addresses a crane scheduling and machine layout problem in a Computer Integrated Manufacturing Environment. A single crane is used to move all the Work-in-Process (WIP) in the system. The overall system objective is to maximize the yield rate subject to the flow time limit of the WIP. The authors formalize the problem, and analytically and empirically show that cyclic scheduling provides a near optimal solution, which is superior to dispatching rules. First, they illustrate the optimality and benefits of cyclic scheduling in a simple environment. Then, for multiple-product problems, the authors show that for a given sequence, finding the minimum cycle time becomes the maximum cost circular network flow problem in a graph. Based on the insights developed, a heuristic for sequencing product types in a cycle is derived that approximately minimizes the cycle time over all sequences. Finally, computational experiments are reported and various assertions made in paper are empirically verified.