A heuristic approach to an interdependent restoration planning and crew routing problem

Abstract

This study proposes efficient solution methods to solve interdependent restoration planning and crew routing problems. The solutions provide reliable restoration plans regardless of the size and structure of disrupted infrastructure networks. We propose a relaxed mixed integer linear programming (MILP) model and two heuristic algorithms to find efficient feasible initial solutions for the restoration routing problem. We also propose a local search heuristic algorithm to find a near-optimal solution using the results obtained from the proposed model. Using electric power, water, and gas infrastructure network instances from Shelby County, TN, the computational results corroborate the efficacy of the mathematical formulation and shows that the heuristic algorithm obtains optimal or near-optimal solutions. In particular, we apply the sequence of the relaxed model, initial solution algorithm, and the local search algorithm for 62 scenarios with different magnitudes of disruption and different numbers of restoration crews. Analyzing the performance of the local search algorithm, we confirm the advantages of using the initial solution algorithm in producing the restoration schedules with reliable and relatively low total restoration time, particularly for large size problems. The observations also reveal how the scattering intensity in the distribution of disrupted locations affects the performance of relaxed models, and consequently, the integrated heuristic.