Optimal policies for stock redistribution in a retail network: Mathematical modeling and algorithmic solution

TL;DR

This paper models and compares centralized and decentralized stock redistribution policies in retail networks, introducing a tailored algorithm to solve large-scale MILP problems based on real industry data.

Contribution

It presents a flexible model for stock redistribution policies and develops a specialized algorithm for efficiently solving large-scale MILPs in retail logistics.

Findings

The model accommodates different redistribution policies.

The custom algorithm outperforms standard solvers on benchmark instances.

Empirical analysis based on real retail data demonstrates practical applicability.

Abstract

We study the problem of stock replenishment and transshipment in the retail industry. We develop a model that can accommodate different policies, including centralized redistribution (replenishment) and decentralized redistribution (lateral transshipments), allowing for direct comparisons between them. We present a numeric analysis in which the benchmark instances stem from the collaboration with a high-end clothing retail company. The underlying model, as usually in the field, is a large-scale mixed-integer linear programming problem. We develop a specific algorithmic procedure to solve this MILP problem and compare its performance with the direct solution via state-of-the-art solvers.