Cluster-first route-second approach for employee shuttle bus routing problem: A large-scale company application

Developing logistics systems is one of the most important strategic steps companies can take toward differentiation in today's global market. In recent years, in addition to the transportation services of the products in the production system, the logistics services of the personnel working within the company have also gained importance. Personnel transportation services are examined within Vehicle Routing Problems (VRP) scope. In this thesis study, school bus routing problems (SBRP), which is a subtype of VRP, are examined and the personnel transportation problem is solved with the SBRP analogy. The personnel transportation problem discussed in the thesis has been solved with the SBRP analogy. Within the scope of the thesis problem, it is aimed to route the transportation between the house of 1361 personnel working in Kartal Carpet& Textile company and the factory. For the identification of personnel addresses and current situation analysis, markings and current situation routing was made with Google Earth Pro's help. Three different vehicle capacity scenarios were created to solve the problem. The problem was solved with the Capacity Vehicle Routing Problem (CVRP) without clustering in the first scenario. Then, P-median, K-means, and P-center clustering methods were applied and optimal results were sought with Cluster-first Route-second (CFRS). The models were run using Python 3.8-Gurobi 9.1.1. In the last part, all the data obtained were analyzed and compared and suggestions were made for future studies.