The capacitated lot-sizing problem with product recovery (CLSP-RM) holds significant importance in reverse logistics but is notoriously complex (NP-hard). In this study, two techniques are introduced to confront this challenge. The first technique entails devising a linear optimization task that eliminates capacity limitations across a wide problem spectrum, yielding a remarkably accurate approximation of the optimal solution (Model A). This adaptable approach presents a potent alternative and holds potential for extension to diverse problem categories owing to its versatile nature. The second technique (Model B) employs a simulation methodology utilizing Halton’s uniform random numbers to address the issue. This randomized production search method sidesteps considerations of production costs, inventory expenditures, and production order when determining production batches . Here are the input data and solution of each instance solved with model A and model B respectively. There are about 4000 instances.
Steps to reproduce
The instances are strictly randomly generated under the assumptions of the paper. All steps that are necessary are described in the paper.