Hierarchical redundancy allocation for multi-level reliability systems employing a bacterial-inspired evolutionary algorithm

A high level of reliability is a crucial factor in many real-world systems. Specifically, redundancy allocation problems (RAPs) have attracted much attention in the last three decades for their comprehensive applications in various engineering systems. RAPs have been proven to be NP-hard and numerous meta-heuristic methods have been proposed to address them. To date, a significant number of successful research endeavors regarding RAPs in single-level systems have been conducted; however, most real complex systems involve multiple levels. Consequently, RAPs on multi-level systems (labeled here as MLRAPs) are deemed to be more realistic and challenging. As such, this paper proposes a bacterial-inspired evolutionary algorithm (BiEA) for addressing MLRAPs. Apart from designing the mutation operation of the canonical bacteria evolutionary algorithm (BEA) to be adaptive to the solution encoding of MLRAPs, two new search operators, dynamic gene transfer and a niching scheme, are introduced in the BEA. As a result, the BiEA is dedicated to MLRAP designs with near-optimal solutions. Case studies on MLRAP designs show the BiEA outperforms current state-of-the-art approaches in regard to two representative experiments.