Redundancy allocation in series-parallel systems under warm standby and active components in repairable subsystems

Abstract

Redundancy allocation is one of the most common approaches to increase the system reliability. In this study, a new model is developed to maximize mean time to failure and to minimize the cost of a system. In general, many researchers are now considering the active redundancy even more than before; however, it is possible for a particular system design to utilize active redundancy and warm-standby redundancy as well. In this model, each subsystem can use both active and warm-standby strategies simultaneously. Moreover, the model allows for component mixing such that components of different types may be used in each subsystem. Thus, the aim of the proposed model is to select the best redundancy strategy, components’ types and levels of redundancy for each subsystem. The simulation and neural network methods are applied considering the structural complexity of the model and repairable components. In order to solve the problem, meta-heuristic of Multi Objective Water Flow algorithm (MOWFA) is proposed and compared to NSGA-II and NRGA. Also, for tuning the meta-heuristics parameters, the Taguchi design of experiments is employed. The algorithms are used to solve 32 test problems and the results are compared. Finally, the results are analyzed and discussed.