Title: Maximum entropy approach to machine repair problem

Authors: Jau-Chuan Ke, Chuen-Horng Lin

Addresses: Department of Applied Statistics, National Taichung Institute of Technology, Taichung 404, Taiwan. ' Graduate School of Computer and Information Technology, National Taichung Institute of Technology, Taichung 404, Taiwan

Abstract: For an M/G/1 machine repair problem, the steady-state probability distributions of the number of down machines are usually found by using the supplementary variable technique and recursive method. We use in this paper the required steady-state availability and the expected number of down machines to compute the maximum entropy approximation for the steady-state probability distributions of the number of down machines in the system, thus avoiding the use of supplementary variable technique. In numerical examples, comparison of the steady-state probability distributions obtained by the exact approach and the approximation procedure are provided to illustrate the effectiveness of maximum entropy approach.

Keywords: bisections; Lagrange multipliers; machine repairs; machinery; maximum entropy; steady-state probability distributions; down machines; supplementary variable techniques; recursive methods; steady-state availability; approximation procedures; service operations; informatics.

DOI: 10.1504/IJSOI.2010.034035

International Journal of Services Operations and Informatics, 2010 Vol.5 No.3, pp.197 - 208

Published online: 06 Jul 2010 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article