Determining the Optimal Correlated Multistate Resource Assignment with Maximal Network Reliability Using a Hybrid GA-TS Algorithm

From a perspective on quality management, network reliability maximization is very important for system supervisors. Despite the numerous studies related to network reliability maximization, no study takes correlated failure into account. This paper focuses on determining the optimal multistate resource assignment to maximize network reliability with correlated failures, where a correlated failure of a multistate resource may be from a large-scale disaster or routine maintenance and thus affects the network reliability. A hybrid algorithm integrating a genetic algorithm (GA) and a tabu search (TS) is developed to solve the addressed problem, in which the network reliability associated with a resource assignment is evaluated in terms of a correlated binomial distribution model and minimal paths. A Taiwan academic network is utilized to demonstrate the computational efficiency of the proposed hybrid algorithm by comparing it with GA and TS.