Elektrikli Araçların Detaylı Enerji Tüketimini Dikkate Alan Dinamik Gezgin Satıcı Problemi
Özet
In recent years, electric vehicles have started to be used frequently in logistics operations, especially in cities, since they cause less damage to the environment compared to conventional vehicles. Instantaneous data from vehicles obtained with the help of the development in vehicle technologies and the progress in information technologies enable to have more efficient and less costly distribution routes. In this study, first, the dynamic travelling salesman problem, which is taken into account with the assumption that vehicle speeds can change dynamically during travel, is addressed. An algorithm based on dynamic programming and the link elimination method have been developed for the problem under the conditions of deterministic battery consumption. The proposed approach has been applied to 90 problems. In 51 of those problems, it has been found to perform better than the Restricted Dynamic Programming Algorithm. The other problem covered by the study is the dynamic travelling salesman problem with uncertain battery consumption, where the battery consumption rate is not precisely known due to the uncertainty of the traffic density. A dynamic programming-based and Prim's algorithm supported heuristic algorithm have been proposed for the problem. With the proposed approach, the purpose is to obtain transportation plans that will aid in lowering urban traffic density, emissions, and range anxiety among electric vehicle drivers in circumstances where traffic density and required battery capacity are both uncertain in advance and change throughout the transportation process. It has been found out that the proposed algorithm outperformed the Restricted Dynamic Programming algorithm in terms of total energy consumption by an average of 6.87%, based on an analysis of 270 large-scale problems. For large-scale routing problems that arise in short-haul freight transportation, distribution plans can be made using either of the decision support models that account for precise energy use.
