A robust optimization approach for humanitarian needs assessment planning under travel time uncertainty

We focus on rapid needs assessment operations conducted immediately after a disaster to identify the urgent needs of the affected community groups, and address the problem of selecting the sites to be visited by the assessment teams during a fixed assessment period and constructing assessment routes under travel time uncertainty. Due to significant uncertainties in post-disaster transportation network conditions, only rough information on travel times may be available during rapid needs assessment planning. We represent uncertain travel times simply by specifying a range of values, and implement robust optimization methods to ensure that each constructed route is feasible for all realizations of the uncertain parameters that lie in a predetermined uncertainty set. We present a tractable robust optimization formulation with a coaxial box uncertainty set due to its advantages in handling uncertainty in our selective assessment routing problem, in which the dimension of the uncertainty (number of arcs traversed) is implicitly determined. To solve the proposed model efficiently, we develop a practical method for evaluating route feasibility with respect to the robust route duration constraints, and embed this feasibility check procedure in a tabu search heuristic. We present computational results to evaluate the effectiveness of our solution method, and illustrate our approach on a case study based on a real-world post-disaster network.