Performance modeling of automated manufacturing systems with unreliable machines and random processing times

Abstract

Reliability plays a very important role in the overall performance of flexible manufacturing systems (FMS). Failure of a single component can easily propagate downtime throughout the entire system, giving rise to inferior and lost products, missed due dates and causing significant losses to the company. Considering that installing an FMS involves a substantial capital investment, it goes without saying that once in place such systems should perform consistently to give an adequate return on the equity that the owners have provided. To secure that objective, FMSs have to be analyzed at their planning, development and implementation stages to ensure reliability. The primary goal of this research has been to develop a methodology which quantifies and includes reliability within the design of flexible manufacturing systems. This is achieved by utilizing an efficient heuristic approach called Group Method of Data Handling (GMDH). GMDH is a self-organization technique that has had numerous successful applications in the areas of environmental, ecological, social systems, and more recently, manufacturing systems. It has proven very useful for modeling complex nonlinear systems with a great number of parameters and a small amount of collected data. The developed performance measure models are capable of predicting the output of FMSs and permitting the user to establish a priori, desired reliability parameter levels within the design and/or operational phases of FMS activity. The proposed approach offers a valuable tool for studying the reliability of FMS during both their design and operation phases.