151. Bounds on the Productivity of Stochastic Production Lines by Probability Masses Fitting
- Author
-
Louvain School of Management - Accounting & Finance, Tancrez, Jean-Sébastien, Semal, Pierre, Chevalier, Philippe, IESM'2007: International Conference on Industrial Engineering and Systems Management, Louvain School of Management - Accounting & Finance, Tancrez, Jean-Sébastien, Semal, Pierre, Chevalier, Philippe, and IESM'2007: International Conference on Industrial Engineering and Systems Management
- Abstract
A modeling method is presented for the analysis of production lines with generally distributed processing times and finite buffers. First, the distributions are discretized by "probability masses fitting". In this paper, we study the "grouping at the beginning" principle, which concentrates the probability mass in a time step at its beginning. Second, the evolution of the line is described by a Markov chain. The method allows to compute refinable bounds on the throughput. We prove that "grouping at the beginning" leads to a better lower bound than "grouping at the end", which have been studied in earlier work. Moreover, we propose some approximations and show that combining both discretization principles brings a significant improvement. Finally, we show how the method performs on examples.
- Published
- 2007