Binary-state line assignment optimization to maximize the reliability of an information network under time and budget constraints

Abstract

Information networks are critical media for various applications of cloud computing and the Internet of Things. How to maintain the system reliability of an information network is an important issue for system supervisors. The system reliability is influenced by the capacities of the lines assigned to the edges that connect the vertices of the information network, where each line should operate to provide a certain capacity or fail. This study aims to maximize the system reliability of an information network by determining the optimal binary-state line assignment under an assignment budget constraint. In contrast to previous studies related to system reliability maximization, this study considers a time threshold in system reliability evaluation. In addition, the information network is regarded as a stochastic information network, as each edge may include several binary-state lines and provide multiple states. A method based on minimal paths and genetic algorithm (GA) with elite replacement is developed to evaluate the system reliability and search for the optimal solution. Several numerical experiments are conducted to show that the GA not only determines the exact solution but also exhibits higher computational efficiency than several well-known meta-heuristic algorithms.