Bi-objective TSP with multiple drones for post-disaster rapid damage assessment
Sprache des Vortragstitels:
Sprache des Tagungstitel:
Rapid and accurate post-disaster damage assessment is critical for humanitarian organizations to plan an effective and efficient disaster relief response. A prior investigation in the affected areas helps humanitarian organizations to be more prepared financially and operationally. On the one hand, the assessment should be performed as fast as possible due to the limited assessment horizon. On the other hand, since visiting all the affected nodes within a limited time may not be possible, the larger explored area of the affected region provides more accurate data for further relief operations. To this end, we consider a case when a truck and drones collaborate to improve the assessment operations exploiting the benefits of using both drone and truck in a sudden-onset disaster setting. In this study, we propose a scenario-based two-stage bi-objective variant of the traveling salesman problem with multiple drones (TSP-MD). The first objective aims at maximizing the number of assessed nodes given their priority scores, whereas the second objective minimizes the assessment completion time. Taking into account the transportation network conditions in a disaster setting as well as the uncertain wind conditions, we consider the truck travel time and the drone travel time as the uncertain parameters. We study the performance of the model for different levels of risk (e.g., expected value and worst-case) to deal with the uncertainties. To solve the proposed models efficiently, we develop a Benders decomposition-based approach. We conduct numerical studies on a set of test instances to evaluate the performance of the proposed approach.