|
Distance minimizing vehicle routing problem for a multi-depot chain retail store: a case study |
|---|---|
| รหัสดีโอไอ | |
| Title | Distance minimizing vehicle routing problem for a multi-depot chain retail store: a case study |
| Creator | Guy Jariyavattanavijit |
| Contributor | Jirachai Buddhakulsomsiri, Advisor |
| Publisher | Thammasat University |
| Publication Year | 2568 |
| Keyword | Vehicle routing problem, Multi-depot routing, Distance minimization, Clustering, Sub-clustering, Capacitated vehicle routing problem, Retail logistics, Excel solver |
| Abstract | The Multi-Depot Vehicle Routing Problem (MDVRP) presents significant challenges in retail logistics when multiple depots must serve a large number of geographically dispersed stores efficiently. This study develops a distance-minimizing decomposition framework for a real-world chain retail network in Bangkok consisting of 3 distribution centers and 148 Tops stores. Because the full-scale MDVRP is too large to solve directly in Microsoft Excel Solver, the problem is addressed through a three-stage framework: depot-store clustering, savings-based sub-clustering, and Capacitated Vehicle Routing Problem (CVRP) optimization. In the first stage, stores are assigned to depots using a nearest-depot rule and then rebalanced using a relative distance ratio criterion, resulting in three balanced depot clusters of 46, 52, and 50 stores. In the second stage, each depot cluster is divided into solver-compatible sub-clusters using Clarke and Wright savings values, demand utilization checks, and a geographic plausibility review. In the third stage, a CVRP mixed-integer linear programming model is solved separately for each sub-cluster and for each delivery day, Thursday and Sunday, using Excel Solver. The resulting routing plan serves the full network with total optimized distances of 2,606.46 km on Thursday using 29 vehicles and 2,891.83 km on Sunday using 32 vehicles. The study shows that a practical cluster-first, sub-cluster-second, route-third framework can make large multi-depot routing problems tractable in a spreadsheet environment while producing implementable distance-minimizing routes. |
