Resilience-driven repair sequencing decision under uncertainty for critical infrastructure systems

Disruptions on critical infrastructure systems (CISs) might affect their normal operation and cause severe social and economic impact. The rapid recovery of post-disaster CISs is crucially important. Based on the authors' previous work on the deterministic formulation of the repair sequencing decision problem of post-disaster CISs under limited repair resources, this paper considers the uncertainty of the repair time of damaged components and proposes a two-stage stochastic model. To solve the stochastic model, a scenario generation and reduction method is first used to generate a limited number of repair time scenarios, and then an efficient enumeration algorithm is proposed for small-scale disruptions and adopted as a module in a heuristic algorithm for large-scale disruptions. Numerical experiments on three systems are conducted to demonstrate the efficiency of the proposed algorithm. Results show that the proposed algorithm can be applied to the recovery decision of large-scale CISs with extensive disruptions. In addition, results show that for most instances, the stochastic solution solved using a limited number of repair time scenarios might be worse than the solution solved from the deterministic model when judging the two solutions by a huge number of repair time scenarios to sufficiently capture the repair time uncertainty.

Automated summary

This paper proposes a two-stage stochastic repair model that considers uncertainty, and solves the recovery decision problem under large-scale and small-scale disruptions using scenario generation and reduction method and an efficient enumeration algorithm. Numerical experiments demonstrate the efficiency of the proposed algorithm and indicate that the stochastic solution might not be as good as the deterministic solution when a large number of repair time scenarios are considered.