Streamlining multi-commodity inventory allocation and redistribution

Abstract

To prevent excess unsold goods caused by market fluctuations, retailers can redistribute surplus commodities among their stores to maximise profits. This paper introduces a novel approach to model the common real-world problem of multi-commodity inventory allocation and redistribution. Our unified approach integrates the two problems and seeks to optimise maximum profit. The study encompasses various factors such as inventory capacity, reallocation constraints, vehicle capacity, time windows for pickup and delivery, and a homogeneous fleet of vehicles. We propose two mixed-integer programming paradigms, the integrated and sequential formulations, along with an improved variable neighborhood search (IVNS) algorithm to solve the problem. Computational results demonstrate the effectiveness of the IVNS algorithm, while further analysis highlights the pros and cons of the two formulation paradigms. Notably, the integrated formulation yields superior solutions at the expense of increased computational time.