OPTIMAL DESIGN OF A MULTI-STATE WEIGHTED SERIES-PARALLEL SYSTEM USING PHYSICAL PROGRAMMING AND GENETIC ALGORITHMS
Wei Li (),
Ming J. Zuo () and
Ramin Moghaddass ()
Additional contact information Wei Li: Citigroup, 33 Canada Square, Canary Wharf, London E14 5LB, UK
Ming J. Zuo: Department of Mechanical Engineering, University of Alberta, Edmonton, Alberta, T6G2G8, Canada
Ramin Moghaddass: Department of Mechanical Engineering, University of Alberta, Edmonton, Alberta, T6G2G8, Canada
In this paper, we report a study of the reliability optimal design of multi-state weighted series-parallel systems. Such a system and its components are capable of assuming a whole range of levels of performance, varying from perfect functioning to complete failure. There is a component utility corresponding to each component state. This system model is more general than the traditional binary series-parallel system model. The so-called component selection reliability optimal design problem which involves selection of components with known reliability characteristics and cost characteristics has been widely studied. However, the problem of determining system cost and system utility based on the relationships between component reliability, cost and utility has not been adequately addressed. We call it optimal component design reliability problem which has been studied in one of our former papers and continued in this paper for the multi-state weighted series-parallel systems. Furthermore, comparing to the traditional single-objective optimization model, the optimization model we proposed in this paper is a multi-objective optimization model which is used to maximize expected system performance utility and system reliability while minimizing investment system cost simultaneously. Genetic algorithm is used to solve the proposed physical programming based optimization model. An example is used to illustrate the flexibility and effectiveness of the proposed approach over the single-objective optimization method.