The recovery of used products and materials has been encouraged for many years, not only by economic reasons but also by environmental reasons. Laws have also been implemented requiring manufacturers to recycle their products. So we are moving from an industrial society highly dissipative in materials, to a society based on Recycling (Mathieux, 2002). In logistical terms, this societal evolution gives rise to an increasing flow of materials from the consumer to producer : the reverse flows. We are interested in the cases where the producer is facing the returns of its own products, in varying quantities. When this return flow is added to the original production flow, it disrupts and complicates the traditional inventory and production management (DeCroix et al., 2005). The objective of this thesis is then to propose some inventory management models that include reverse flows. On these models, wi will look for e±cient strategies of supplying inventories which take into account the random return flows, in order to reduce overall system cost. Furthermore, and beyond the search for these procure- ment strategies, an important part of our contributions lies in the numerical studies which will be conducted on our models to test the validity, usefulness or relevance of some assumptions.
