Abstract:

Today, the number of natural disasters is increasing and occurring more frequently, and these disasters deeply affect human life. Dealing with the devastation caused by natural disasters such as earthquakes, floods, and pandemics is quite challenging. The earthquake in Turkey on February 6 affected 11 provinces and affecting approximately 14 million people. After earthquakes, transportation infrastructure like roads, bridges, tunnels, and railways can become non-functional, making it challenging to quickly determine alternative routes. Vehicle routing problems (VRP) approaches can offer solutions for post-earthquake relief distribution activities. VRP is a combinatorial optimization and integer programming problem that optimizes a vehicle fleet to serve many customers. The Vehicle Routing Problem with Time Window (VRPTW) aims to determine routes with the lowest cost under specific time and capacity constraints. In this study, a Clustering-Based Clonal Selection Algorithm (CSA) is proposed for VRPTW. The initial solution set of the algorithm has been improved using K-means and K-means++ algorithms, and then results for VRPTW have been obtained with CSA. Experiments were conducted on the Solomon C1 and R1 datasets frequently used in the literature for testing VRP algorithms, and results were obtained for various problems. According to the experimental results, obtaining an initial solution with the clustering algorithm improved the results of the CSA and prevented the CSA from getting stuck in a local optimum.

Keywords: