Scheduling periodic customer visits for a traveling salesperson |
| |
Authors: | Michael Polacek Karl F. Doerner Richard F. Hartl Guenter Kiechle Marc Reimann |
| |
Affiliation: | 1. University of Vienna, Department of Management Science, Bruenner Strasse 72, 1210 Vienna, Austria;2. Salzburg Research Forschungsgesellschaft mbH, Jakob Haringer Strasse 5/III, 5020 Salzburg, Austria;3. Institute for Operations Research, ETH Zurich, Clausiusstrasse 47, 8092 Zurich, Switzerland |
| |
Abstract: | The problem considered in this paper deals with determining daily routes for a traveling salesperson who provides customers in Upper Austria with product range information of a large, global food wholesaler. Each customer has to be visited at least once a year, with some customers requiring up to one visit per month. Further, some customers may not be visited each day of the week. Our decision support system uses a commercial GIS software to extract customer data for input into the optimization procedure and to visualize the results obtained by the algorithm. The optimization approach is based on the variable neighborhood search algorithm which assigns customers to days and determines routes for the salesperson for each day with the primary objective to minimize the total travel time of the salesperson. Another objective studied is to minimize the number of days needed by the salesperson to visit all customers in a given month. Further we analyze the effects of changes in the business environment like increases in the amount or flexibility of the salesperson’s working time and variations in the possible days for customer visits. Finally, we enrich the objective function by considering periodicity requirements for customer visits. Specifically, we penalize irregular schedules, where the time between two successive customer visits varies. |
| |
Keywords: | Periodic Traveling salesperson problem Variable neighborhood search Real world problem |
本文献已被 ScienceDirect 等数据库收录! |
|