Optimization of reliability redundancy allocation problem with component mixing and strategy selection for subsystems

Abstract

The reliability-redundancy allocation problem (RRAP) is usually formulated based on two main assumptions, (1) using homogeneous components, and (2) having a predetermined redundancy strategy for each subsystem. In this article, both these restrictive assumptions are relaxed to create a more realistic problem. This leads to a new formulation of the problem with heterogeneous components. Moreover, an optimal strategy is developed for each, either active or cold standby, subsystem based on the mathematical model proposed herein. The reliability of standby subsystems is calculated using an exact Markov model. The proposed approach is tested on three different benchmark test problems. For solving the mathematical models, a parallel stochastic fractal search algorithm is developed. Comparison of the results obtained with similar studies previously reported demonstrates the advantages of the new RRAP model.