InstancesCLSP-RM (Paper: https://www.preprints.org/manuscript/202304.0242/v4)

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. (The test instances and solutions are available here: https://www.preprints.org/manuscript/202304.0242/v4 )