Vehicle routing for relief supplies distribution under uncertainty of natural disaster

Abstract

Flooding is a severe and widespread natural disaster that significantly impacts the environment and infrastructure and adversely affects human lives. In affected areas, transportation systems are often disrupted, leading to critical shortages of essentials such as food and water. Rapid and flexible delivery of relief goods via vehicles is crucial for sustaining life and aiding community recovery. This paper presents a new model called the Vehicle Routing Problem for Relief Item Distribution under Flood Uncertainty (VRP-RIDFU), which focuses on optimizing the speed of route creation and minimizing the waiting times for aid delivery in flood conditions. The Genetic Algorithm (GA) manages uncertainties and effectively solves NP-Hard problems. This model comprises a dual-population strategy: a random and enhanced population. The latter is designed to handle uncertainties, assessed by evaluating the expected performance of routes based on waiting times and flood risks. The Population Sizing Module (PSM) has been developed to automatically adjust the population size based on the dispersion of affected nodes, which is assessed using standard deviation. The Complete Subtour Order Crossover (CSOX) operator is introduced to enhance the quality of routes and expedite convergence. The model's performance has been validated through simulated flood scenarios that exhibit a variety of uncertainties in road conditions. Prioritizing waiting times over travel times in routing decisions has proven effective. This model has been tested with 20 standard CVRP problem sets, each with different numbers and distribution patterns of nodes. The test results surpass those of the shortest routes, which serve as the benchmark for optimal solutions. The outcomes affirm the model's capability to generate high-quality vehicle routing plans in all tested scenarios rapidly.