Abstract (eng)
Most vehicle routing models assume constant travel times throughout the whole planning horizon. In reality, however, travel times vary during the day. This is especially true for urban areas where daily traffic congestion leads to longer travel times. The time-dependent vehicle routing problem (TDVRP) takes this aspect into account by assuming that travel times depend on the time of the day.
This diploma thesis gives an overview of the TDVRP and presents the results of an experimental study.
The first part introduces the VRP and different solution methods. This is followed by a detailed description of the TDVRP.
The second part of the thesis presents an algorithm based on tabu search to solve the capacitated VRP (CVRP). Afterwards, the best solutions of the CVRP are evaluated with five time-dependent scenarios, each representing a different degree of time-dependency. Compared to the original CVRP results, the total costs increase significantly and several routes become infeasible. In the next step, the original algorithm is adapted to solve the TD-CVRP. It is shown that the total costs can be improved when assuming time-dependent travel times. The improvement is higher, the higher the degree of time-dependency. Furthermore, the new solutions satisfy all tour length constraints.