Abstract
There has been a growing interest in how to find the optimal delivery route for efficient usage of corporate delivery vehicles. Contrary to this, little attention has been paid to generalized link costs between delivery points. The costs should be given as travel times in finding the optimal route, because distributions are considered with time constraints especially in urban areas where heavy traffic is encountered.
How to estimate the travel time becomes the important issue. Most of existing studies on travel time estimation are based on data from vehicle monitoring equipment. This kind of estimation, however, seems restrictive for urban deliveries because some areas may not be monitored.
In this study we first built an estimation model based on actual travel data obtained from a delivery of machine tools. By using this model, we gained insights for factors which influence on the estimation of travel time. For example the following factors mostly affect the time estimation : the numbers of right and the distance of the main street in the evening rush hour.
Our model is well built, however has an issue that estimation parameters produced greater error for long distance travel.
