STATIC ONE-TO-ONE MULTI-VEHICLE PICKUP AND DELIVERY PROBLEM WITH MULTIPLE DEPOTS

Abstract

The research describes a static case of the pickup and delivery problem with multiple depots and proposes a solution approach that forces a combination of requests in a particular route. To solve the problem, the model is implemented using General Algebraic Modelling System (GAMS) with CPLEX as the solver. Euclidean distances are considered as the total cost between nodes. Instances based on existing studies are used for computational experiments. Best-bound search and depth-first search methods of the branch-and-bound algorithm are also compared. Results of the trials implemented in GAMS are presented in the study. The optimal solutions found with routes having combined requests show significant distance reduction compared to the typical individual service of one request per route. Results also show that as the number of allowed requests in one route is increased, the effect of the number and variation of depots also decreases. The main contribution of this research are applications in GAMS that can solve the static case of the messenger problem in a multi-depot variation, a setup with few available literature. The study can be beneficial for logistics companies especially those of which specializing in pickup and delivery services.