Partial Network Recovery to Maximize Traffic Demand

A natural disaster will cause large-scale failures and damage many network components, which in turn affect users?traffic. This paper addresses the problem of selecting a subset of the failed components to repair after massive failures so as to meet the given traffic demand as much as possible with limited recovery resources (e.g., workforce). We formulate the problem using mixed integer programming (MIP), and then propose an effective heuristic algorithm to solve the MIP. Simulations are performed to evaluate and compare the performance of the MIP and heuristics.