Use statistical analysis to approximate integrated order batching problem

TL;DR

This paper introduces a statistical analysis-based approach to approximate the integrated order batching and packing problem, aiming to reduce warehouse costs through a novel reformulation and heuristic methods.

Contribution

It presents a new Max Correlation Reformulation framework and a heuristic approach for approximating the complex order batching and packing problem.

Findings

Outperforms existing methods in numerical experiments

Significantly reduces warehouse picking and packing costs

Provides a novel statistical approximation framework

Abstract

Order picking and order packing entail retrieving items from storage and packaging them according to customer requests. These activities have always been the main concerns of the companies in reducing warehouse management costs. This paper proposes and investigates the Order Batching and Order Packing Problem, which considers these activities jointly. The authors propose a novel statistic-based framework, namely, the Max Correlation Reformulation problem, to find an approximation mixed-integer programming model. An approximation model is found within this framework in two phases. A lower dimension model is firstly proposed. Efforts are then made to increase its correlation coefficient with the original formulation. Finally, a powerful pairs swapping heuristics is combined with the approximation model. Numerical experiments show that this newly found approach outperforms the mainstream methods from the literature. It is demonstrated that this proposed method could significantly reduce the cost of picking and packing operations in a warehouse.