Robust design of maintenance systems for a manufacturing system - a computer simulation and experimental-design oriented approach

This paper presents a synthetic decision procedure to determine robust maintenance polices for a robust manufacturing system under a fixed maintenance operation budget. It integrates simulation, (fractional) factorial design, Taguchi design noise array, response surface methods, and nonlinear programming to avoid the intractable problem of complicated mathematical models of maintenance systems which are usually modeled through queuing networks based on many unreasonable and impractical assumptions. The results show that the approach not only provides a new way to find satisfactory maintenance polices for complicated manufacturing systems, but also provides a guide for the design of a robust system under maximum performance function and required constraints.