Optimasi Capacitated Vehicle Routing Problem (CVRP) Menggunakan Kombinasi Algoritmavariant Sweep dan Particle Swarm Optimization (PSO)

Abstract

Distribution route planning is an important aspect of logistics systems, especially for subsidized fertilizers, which must reach farmers efficiently and on time. In Bangsalsari District, Jember Regency, subsidized fertilizers are distributed from a depot to several kios with varying demand, and the vehicles used have limited capacity. In this study, the distributor CV Hidup Baru serves 29 fertilizer kios located in Bangsalsari District. This condition can be modeled as a Capacitated Vehicle Routing Problem (CVRP), where the objective is to determine the shortest distribution route while considering vehicle capacity constraints. This study aims to determine the optimal distribution route for subsidized fertilizer delivery in Bangsalsari District using a combination of the Variant Sweep Algorithm and Particle Swarm Optimization (PSO). The Variant Sweep algorithm clusters kios based on their polar angles relative to the depot. In contrast, the PSO algorithm is applied to determine the optimal visiting order within each cluster. The research uses primary data comprising kios coordinates, fertilizer demand, depot locations, and vehicle capacity, obtained from the Department of Food Crops, Horticulture, and Plantation of Jember Regency. The model is implemented in Python, and distance calculations are obtained via the OpenRouteService API. The results show that the Variant Sweep algorithm generates ten different cluster arrangements, each consisting of eleven clusters, from the 29 kios based on vehicle capacity constraints. The PSO algorithm then optimizes the visiting sequence within each cluster. The best solution is obtained from the eigth cluster arrangement with a total distribution distance of 176.57 km.