Abstract:

Vehicle Rotation Problem (ARP) is one of the problems that can be seen in the distribution field. Vehicles starting to tour from a central warehouse in the problem, visit customers and return to the warehouse again. There are many types of ARP. The fuel distribution also offers a different version of the ARP, called the Comparted Vehicle Rotation Problem (KARP). The fuel products that should not be mixed in the KARP are transported in different compartiments. As in ARP, the solutions for KARP are usually divided into two categories: route creation intuitions and improvement intuitions. Although there is a broad literature about ARP related to both intuitive categories, literature for KARP is limited. The aim of this study is to propose a starting solution to the fuel distribution problem in Turkey. There are variables that make the problem KARP, each customer more than once, comparison, tolerance. Saving Algorithm (TA) is a starting solution for the problem, integrated with a loading algorithm.  The recommended method has been tested on test problems and has given acceptable results compared to literature. The proposed method has solved the real-life fuel distribution problem and significant cost improvement has been made. The recommended method can be used both as a starting solution in the KARP problems and as a practical application in the distribution of fuel in Turkey.

Abstract:

One of the well known problems in distribution is the Vehicle Routing Problem (VRP). In this problem, vehicles start their tours from a depot, they visit the customers and finish their tours in the depot again. There are several variants of VRP. Fuel distribution provides a different kind of VRP which is called as Multi Compartment Vehicle Routing Problem (MCVRP). In MCRVP vehicles have compartments and incompatible petroleum products are delivered in different compartments of a vehicle. For both of the problems solution methods can be classified into two categories; route construction heuristics and improvemenet heuristics. While there is a significant literature on VRP regarding these heuristics, the literature on MCVRP is much more limited. The aim of this study is to offer a construction heuristics for a fuel distribution problem in Turkey. The problem has some features such as compartments, visiting customers more than once and tolerance which turns it into a MCVRP problem. Savings Algorithm (SA) and a loading algorithm is integrated to develop a construction heuristics for the problem. Proposed heuristics is tested on test problems and acceptable results obtained when compared with the literature. A real life fuel distribution problem is solved with the propsed heuristics and significant cost reduction has been achieved. The proposed method can be used as a construction method in MCVRP and as a practical solution for real life problems in fuel distribution in Turkey.

Keywords: