Development of a demand uncertainty-based model to estimate network reliability

Abstract

This paper provides a discussion on network reliability, which is a subset of resilience. Past research shows how better understanding reliability and identifying ways to improve it can help a system’s response to a disaster, both in the annealing and recovery stages of the four-step resiliency cycle. This study seeks to develop a reliability model that also considers uncertainty. The model estimates reliability for three performance functions (travel time, flow, and consumer surplus), and accounts for uncertainties in both reductions in link-capacity and consumer demand in response to a perturbation/disaster. The model is coded in MATLAB and a numerical example is shown, complete with sensitivity and uncertainty analyses. The results show that this model provides accurate predictions of the expected function of the system. Particularly in comparison to those of the deterministic scenarios, the scenarios incorporating uncertainty yield far superior results.