Screening of optimal structure among large-scale multi-state weighted k-out-of-n systems considering reliability evaluation

Keen interest in the development of multi-state weighted k-out-of-n systems (MSWKNS) has been currently observed. In real-life cases, considering that multiple different MSWKNS structures are provided beforehand, system designers need to select an optimal one among numbers of different structure designs. Thus, screening of the optimal MSWKNS structure is meaningful and critical. Moreover, the system reliability becomes an increasingly important factor that should be assessed in structure screening. However, when facing a large-scale MSWKNS whose number of states is enormous, exact reliability evaluation can be rather complicated. Meanwhile, due to the large number of system structure designs, the computational burden of screening can be very huge. To effectively select the optimal structure among enormous MSWKNSs under reliability constraint, an optimal structure screening method named as ordinal structure screening (OSS) is proposed. Ordinal optimization (OO) algorithm is adopted with a novel utilization of continuization discretization approximation (CDA) method. To shorten the reliability evaluation time, CDA is employed to approximate reliabilities. Based on the approximate reliabilities, OO is applied to reduce the number of system structures to be evaluated, which accelerates the screening process. Illustrative results show that the proposed method appeals in improving computational efficiency with satisfactory accuracy performance.