All-Terminal Network Reliability Estimation with Monte Carlo and Deep Neural Networks
Conference Paper
Overview
Research
Additional Document Info
Other
View All
Overview
abstract
All-terminal network reliability is a crucial feature as it provides a holistic measure for critical infrastructures such as transportation, computer, and communication networks. Fast and accurate network reliability estimation can help to prevent mishaps. Exact all-terminal calculation is an NP-hard and computationally expensive problem, which has led to the development of approximate methods based on artificial neural networks (ANNs). Although there have been some successful research studies on all-terminal reliability estimation using ANNs, they usually require the exact reliability to train the ANNs. Exact reliability calculation is time consuming and might not be practical for networks with more than ten nodes. Moreover, perfect nodes have been primarily assumed, considering that only links can fail. However, in reality, both kind of components, i.e., nodes and links, may fail. Due to the complexity of the problem, and as an alternative to get fast estimations of the reliability of a larger network, an integration of Monte Carlo and Deep Neural Networks (DNNs) is proposed. The proposed Monte Carlo-based algorithm can provide estimation of the network reliability for given nodes and links reliability values. Nevertheless, even this algorithm might not be practical for real-time applications To speed-up the calculation, a DNN is integrated into the framework, thus enabling the estimation of network reliability for given link and node reliability values.
