Non-stationary ordering policies for multi-item inventory systems subject to a single resource constraint and quantity discounts

The multi-item inventory problem with a single resource constraint and quantity discounts is studied in this paper. Known solution approaches to this problem, both heuristic and enumerative, assume that ordering policies are stationary. A new heuristic approach is presented here for generating non-stationary ordering policies, with order quantities that can vary over time. Computational results for comparing these methods are given. The non-stationary approach is shown in these experiments to outperform heuristic stationary approaches, and an enumerative stationary approach when the non-stationary approach is suitably initialized.