| |
| In this paper, we develop a method of analyzing a serial production line with single part-type and multiple parallel-machine workstations, which has not been paid much attention to in existing literature. At first, we analyze a short line with one parallel-machine workstation and two buffers by constructing its probability balance equation based on a Markov chain model, and evaluating its performances such as throughput and work-in-process level. Then a decomposition-based algorithm is developed for the analysis of a long line with multiple parallel-machine workstations by using the result of short line analysis as building block. Our method is different from traditional decomposition-based method in that our decomposition is workstation-centered while the traditional one is buffer-centered. Numerical experiments shows that result of performance evaluation by using the proposed method is quite close to those obtained by simulation, and our method is much less time-consuming. |
|
| Keywords:tandem production line, Markov chain, probability balance equation, decomposition-based algorithm. |
|
