Sequencing post-disruption concurrent restoration via a network flow approach

Abstract

Infrastructure networks play pivotal roles by facilitating social interactions and economic activities. Nonetheless, these networks are susceptible to disruptive influences such as natural disasters. In response to such disruptions, decision-makers must undertake efficient allocation of recovery projects to restore to a pre-disruption state. With simultaneous project restorations, we consider the concurrent restoration process as a discrete dynamical system and construct a restoration network (R-Net), where a unique project allocation is represented as a path from the source node to the sink node. Upon the R-Net, the restoration sequence along with project allocation is optimized by solving a minimum-cost flow problem. Several numerical examples are presented to demonstrate the effectiveness of the network flow approach, and the findings indicate that an increase in the number of recovery projects and a decrease in recovery resources lead to a notable expansion of the R-Net scale.