A Hybrid Sweep-Nearest Neighbor-Tabu Search Approach for CVRP in FMCG Route Distribution

Abstract

This study addresses the Capacitated Vehicle Routing Problem (CVRP) in the distribution of Fast-Moving Consumer Goods (FMCG) by proposing a hybrid approach that combines the Sweep algorithm, Nearest Neighbor (NN) method, and Tabu Search (TS) algorithm. The objective is to satisfy consumer demand and vehicle capacity restrictions while minimizing the overall journey distance. The Sweep algorithm is used to cluster customers based on polar coordinates, the NN method determines initial delivery routes within each cluster, and TS refines those routes to find near-optimal solutions. Implemented on a real-world dataset of 248 stores in Malang, the proposed hybrid method achieved significant reductions in the number of clusters and total travel distance compared to conventional approaches. Results show that the Sweep algorithm successfully reduced the number of delivery clusters from 26 to 18, achieving a 30.77% reduction in grouping efficiency. Using the Nearest Neighbor method, the total route distance was 2,191.08 km. Further optimization with Tabu Search reduced the Distance to 2141.31 km. Compared to the conventional method, which is 2345.90 km, the hybrid approach resulted in an 8.72% improvement in route efficiency. These findings demonstrate that the integrated method is effective for large-scale distribution problems under capacity constraints. The hybrid method offers a practical and computationally efficient solution for large-scale FMCG distribution networks.