An approach for reliability optimization of a multi-state centralized network

•A star configuration network under different activation strategies is studied.•All nodes can be consist of more than one multi-state heterogeneous components.•An algorithm based on the UGF method is developed for reliability evaluation.•A real case study is investigated to show the procedure of the proposed algorithm. Modern technological advances require unprecedented reliability improvements in engineering systems. This is especially true for the numerous complex structures used as infrastructure in modern industries such as communication and power systems. The structures of systems used in these industries are generally designed as centralized networks, mainly configured in a star structure, in which the end nodes are physically connected to a central one. The principal objective of the current study is to develop a novel approach for reliability analysis and design optimization of such networks, in which all nodes might consist of more than one multi-state heterogeneous component. The performance of each node will then be evaluated as affected by different activation strategies with switch failure also considered. Moreover, it is possible for both central and end nodes in star networks to share their surplus performance, in which case transmission losses must be considered. The universal generating function (UGF) method will be exploited to develop the relevant model. Another aspect of this study involves investigating the optimal component allocation, corresponding activation strategies, and resulting optimal system reliability. Finally, the model will be validated using a numerical example and a real case study.