Modeling and evaluation method for resilience analysis of multi-state networks

Multi-state networks which not only be connected but also function at a certain performance level exist widely in real-world engineering systems. Resilience provides a new approach for the design and analysis of networks that enhance the ability of systems to withstand and bounce back from disruptive events. Efficient and accurate evaluation method is essential for analyzing and enhancing the resilience of the multi-state networks. This research focuses on the resilience modeling and evaluation method for multi-state networks, in which the states of the components and network follow certain probability distributions. First, a multi-state network performance metric associated with the state distribution of components was proposed. Then, we proposed a semi-Markov-based resilience model to describe the transition of the component state distribution. Third, we proposed a fast-repeated algorithm to enhance resilience evaluation efficiency for multi-state networks such that managers can make timely decisions when disruptive events occur. Finally, a case study is conducted, where the optimal recovery strategy was obtained based on the proposed method. Compared to the distance-based recovery strategy and reliability-based recovery strategy, resilience-based method offers more flexible and timely strategies considering different mission requirements and/or preferences for the design, operation and recovery of multistate networks.