Cluster-first route-second approach for multimodal personnel transportation: Gaziray suburban case

Efficient planning of employee transportation in urban transportation is of great importance in terms of reducing operating costs and contributing to sustainable transportation goals. In this thesis, the employee transportation problem for 3,197 employees working at a factory operating in Gaziantep is addressed within the framework of the Capacity Vehicle Routing Problem (CVRP). Using the employees' location information, an asymmetric distance matrix was created using the Open Source Routing Machine (OSRM) infrastructure, and the routes for 189 vehicles were obtained via Google OR-Tools using the PATH-CHEAPEST-ARC and GUIDED-LOCAL-SEARCH algorithms. Subsequently, the Gaziray suburban line was integrated into the model. The distances from employees' homes to Gaziray stops were examined according to two different walking distance scenarios: 500 m and 750 m. Employees who were not within a 500 m or 750 m walking distance of Gaziray stops were clustered using the Set Covering model. Each cluster should have a maximum capacity of 17 people, and the walking distance within the cluster should be 500 m in the first scenario and 750 m in the second scenario. After the clustering process, the MD-CVRP model was applied to ensure that the service vehicles reach the nearest Gaziray stations after stopping at the cluster centers.