An Optimization Approach to the Ordering Phase of an Attended Home Delivery Service

TL;DR

This paper presents a new optimization approach for the ordering phase in attended home delivery systems, focusing on dynamically offering delivery windows efficiently during online grocery order placement.

Contribution

It introduces a hybrid heuristic and MILP-based method to determine feasible delivery windows in real-time, improving the efficiency of the ordering process in AHD systems.

Findings

The proposed methods efficiently determine feasible delivery windows.

Hybrid approach outperforms purely heuristic methods in computational experiments.

Approaches are applicable to real-world online grocery delivery scenarios.

Abstract

Attended Home Delivery (AHD) systems are used whenever a supplying company offers online shopping services that require that customers must be present when their deliveries arrive. Therefore, the supplying company and the customer must both agree on a time window, which ideally is rather short, during which delivery is guaranteed. Typically, a capacitated Vehicle Routing Problem with Time Windows forms the underlying optimization problem of the AHD system. In this work, we consider an AHD system that runs the online grocery shopping service of an international grocery retailer. The ordering phase, during which customers place their orders through the web service, is the computationally most challenging part of the AHD system. The delivery schedule must be built dynamically as new orders are placed. We propose a solution approach that allows to (non-stochastically) determine which delivery time windows can be offered to potential customers. We split the computations of the ordering phase into four key steps. For performing these basic steps we suggest both a heuristic approach and a hybrid approach employing mixed-integer linear programs. In an experimental evaluation we demonstrate the efficiency of our approaches.