Article ID: | iaor1989592 |
Country: | United States |
Volume: | 23B |
Issue: | 4 |
Start Page Number: | 271 |
End Page Number: | 281 |
Publication Date: | Aug 1989 |
Journal: | Transportation Research. Part B: Methodological |
Authors: | Laporte Gilbert, Mercure Hlne, Chapleau Suzanne, Landry Philippe-Eric |
Keywords: | programming: travelling salesman |
Mailbox collection is a complex and costly operation in most postal services. The authors recently undertook a study aimed at improving this operation at the Canada Post Corporation (CPC). This paper describes an improvement to the clustering and routing algorithm currently in use. Various traffic and union rules affecting mailbox collection in an urban context are first described and their effect on shortest distance computations is explained. This is followed by a description of the clustering and routing algorithm currently used at CPC. A model and a generalized travelling salesman algorithm that take the various rules into account are then presented. It is then shown how the CPC routes can sometimes be improved by using the proposed algorithm as a postoptimizer.