Volume 22, Issue 3 (9-2025) jor 2025, 22(3): 69-99 | Back to browse issues page

Salehi Sarbijan M. Shared Customers in the Routing Problems with Feeders for Non-Similar Goods Using Particle Swarm Optimization Algorithm with Adaptive Learning Strategy and Dynamic Inertia Weight. jor 2025; 22 (3) :69-99
URL: http://jamlu.liau.ac.ir/article-1-2313-en.html

Shared Customers in the Routing Problems with Feeders for Non-Similar Goods Using Particle Swarm Optimization Algorithm with Adaptive Learning Strategy and Dynamic Inertia Weight

M. Salehi Sarbijan ^*

Assistant Professor, Department of Mechanical Engineering, Faculty of Engineering, University of Zabol, Zabol, Iran , m.salehisarbijan@uoz.ac.ir

Abstract: (44 Views)

The diversity of customers’ geographical locations forces companies’ delivery systems to use longer distances and more vehicles, which leads to congestion and crowding in urban transportation networks, air pollution, long delays in daily travel times, and increased fuel consumption. Therefore, examining the existing challenges in vehicle and customer routing in the supply chain is of great importance. The feeder vehicle routing problem (FVRP) helps reduce distribution costs by utilizing heterogeneous vehicles and the feeder approach between small and large vehicles. Additionally, the collaboration mechanism between companies to serve shared customers leads to significant savings in the number of vehicles, distance traveled, and maximization of capacity utilization. The aim of the present study is to develop the collaborative vehicle routing problem with feeders (CFVRP) within the framework of customer demand sharing and different goods, with the goal of minimizing operational costs. After modeling the CFVRP using the mixed integer linear programming (MILP) model, the standard PSO, PSO with dynamic inertia weight (WPSO), PSO with adaptive learning strategy (LAPSO), and PSO with adaptive learning strategy and dynamic inertia weight (LAWPSO) algorithms were evaluated to solve it. The experiments were conducted on two sets of small instances (10 to 20 customers) and medium - large instances (50 to 290 customers). The results showed that in small instances, the CPLEX model provided optimal solutions with a zero relative gap due to the use of exact methods. In contrast, the LAWPSO algorithm demonstrated superior performance in both instance categories in terms of solution quality and computational efficiency compared to PSO, LAPSO, and WPSO. This superiority was particularly evident in medium and large problems, which require managing the complexities arising from a large number of customers and diverse demands.

Keywords: Collaborative Network, Feeder Vehicle Routing Problem, Different Goods, Mixed-Integer Linear Programming, Particle Swarm Optimization With Adaptive Learning Strategy And Dynamic Inertia Weight

Type of Study: Research | Subject: Special
Received: 2025/05/8 | Accepted: 2025/07/28

Send email to the article author

Rights and permissions
	This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.