Large scale capacitated vehicle routing problem: A case study of a convenience store chain

Abstract

The Capacitated Vehicle Routing Problem (CVRP) is a significant logistics and management NP-hard problem in the industrial sector. The CVRP appears in more large-scale and complicated scenarios as the urban planning sector develops. In CVRP, the primary objective is to minimize the total cost or distance travelled to fulfill the FXVWRPHUV¶ GHPDQG by the vehicles while accommodating each vehicle's capacity. The cluster-first route-second method is one of the techniques utilized to solve the CVRP. It involves arranging consumers into a few clusters, with each cluster being handled by a single vehicle. Once clusters are assigned, a route identifying the optimal order to visit customers inside each cluster is established. The main purpose of this study is to propose two algorithms in the clustering phase, namely Modified Fuzzy C-means with K-means for generating initial centroids and K-means MHHO, to large-scale problems DQG D FDVH VWXG\ LQYROYLQJ WKH GLVWULEXWLRQ URXWH IRU 0LQLVWRS¶ '& 6SHFLILFDOO\ DQ FCM-based technique is given to solve the clustering phase at a lower cost through the initialization of centroids. Also, we attempt to build the K-means MHHO by combining the MHHO algorithm with the basic K-means clustering algorithm and introducing a capacity limit. After the clustering phase, the routing phase of the TSP issue is handled using a modified Ant Colony Optimization technique. Two proposed algorithms are evaluated thorough comparisons and analysis on many benchmark issues, particularly for large-scale datasets, then resolving the case study.