Multi-depot vehicle routing problem for large scale disaster relief in drought scenarios: The case of the Brazilian northeast region

Abstract Emergency water trucking (EWT) is the reactive response to address droughts most adopted by developing countries. Consequently, as droughts are becoming more frequent, severe, and longer due to climate change, there has been an increasing demand for EWT. However, when applied on a large scale to geographically dispersed rural populations, EWT requires the use of a significant vehicle fleet, leading to higher operational costs. It thus becomes crucial to propose solutions that guarantee EWT's financial and technical viability. Therefore, this paper proposes a two-step procedure to address the problem of large-scale water distribution to drought-affected populations. In step one, we use the transportation problem to assign a set of demand points to water sources, and, in step two, we model the problem as Capacitated Vehicle Routing Problems. Moreover, we also explore in this paper possibilities of using a new hybridizing Ant Colony Optimization metaheuristic with Random Variable Neighborhood Descent (MACS-RVND) to search for efficient solutions, in terms of the total distance and number of vehicles used, for large scale EWT. Hence, in the second step of the developed procedure, results from the MACS-RVND are compared with those obtained by Clarke and Wright heuristics with 2-opt, and with results from an Adaptive Large Neighborhood Search approach proposed by Erdogan. The procedure was then applied to a real water distribution case in the Brazilian semi-arid region. Results are compared with the actual procedures adopted in Brazilian water distribution, showing that the developed procedure can promote more efficient, economic, and equitable water distribution.