Sensitivity analysis of failure-prone flexible manufacturing systems

Performance evaluation of failure-prone manufacturing systems is of significant practical interest. In this paper, the authors consider the parametric sensitivity analysis of manufacturing systems producing multiple part types. This involves transient analysis of the Markov model of the system and computation of the derivative of the cumulative throughput with respect to the varying parameters. Such an analysis would be useful in studying the effects of various not exactly known parameters on the system performance, identifying the bottlenecks and hot spots and in system optimization studies. In this paper, the authors derive expressions for the sensitivity of the first two moments of the throughput-related performability with respect to failure and repair rates of various subsystems such as machines, guided vehicles, robots, etc. The authors also present two examples to illustrate the theoretical results of this paper.<<ETX>>