Article ID: | iaor19982594 |
Country: | United States |
Volume: | 8 |
Issue: | 3 |
Start Page Number: | 274 |
End Page Number: | 288 |
Publication Date: | Jun 1996 |
Journal: | INFORMS Journal On Computing |
Authors: | Bitran Gabriel R., Mondschein Susana V. |
Keywords: | programming: dynamic, heuristics, marketing, retailing |
One of the main features of the catalog sales industry is the ability to influence demand through the mailing policies. This characteristic forces catalog managers to define coordinated strategies between mailing and inventory decisions. In this paper we study inventory policies for each item promoted in the catalog, taking into account the mailing policies. For this purpose, we analyze two different approaches: the global and hierarchical models. Both approaches consider the dynamic and stochastic nature of the problem. Special emphasis is given to the development of a feedback strategy between the tactical and operational levels, in the hierarchical approach, to avoid ‘infeasibilities’ induced mainly by inventory imbalances. Given the complexity of the formulations, we develop heuristics based on the optimal solutions of particular problems. Preliminary computational results show that these heuristics have a satisfactory performance when compared with an upper bound derived for the problem. These results also show that the hierarchical approach has an excellent performance when one feedback from the operational to the tactical level is allowed. This interaction mimics what is usually observed in practice.