A stochastic programming approach for designing supply loops

A stochastic programming model for designing logistics networks integrating reverse logistics into current supply chains is proposed in this paper. It aims at evaluating impacts of randomness related to recovery, processing and demand volumes on the design decisions. These decisions deal with the location of service and processing centres and warehouses, regarding processed products with reuse potential, the definition of mission of sites and, consequently, the product accessibility. Flows of recovered products may be directed toward one or a number of processing alternatives, according to the different states of the recovered products as well as to the network conditions, which relate to recovery and demand volumes with respect to capacity constraints and operating costs. Notably, recovered products may be repaired, disassembled for part refurbishing or disposed. Such products are indicated here as valorized products and represent an economical supply alternative, which meets lower quality standards in comparison with new products. Portions of needs fulfilled by valorized products are defined according to the requirements of end-users, as well as management policies and strategies. The model aims at improving valorized product accessibility, while reducing the total operating costs of such a network. A heuristics based on the sample average approximation, involving the Monte Carlo sampling methods, is proposed to solve the problem. (C) 2008 Elsevier B.V. All rights reserved.