A Short-Term Predict-Then-Cluster Framework for Meal Delivery Services

TL;DR

This paper develops a short-term predict-then-cluster framework to improve on-demand meal delivery operations by forecasting zone-level demand and forming dynamic, constraint-aware clusters. It leverages lagged-dependent ensemble-learning models for point and distributional forecasts, and introduces Constrained K-Means (CKMC) and Contiguity Constrained Hierarchical Clustering with Iterative Constraint Enforcement (CCHC-ICE) to generate operationally viable clusters. Case studies in Europe and Taiwan show that distributional forecasts and lagged features improve clustering quality and operational insights, while a fleet-rebalancing simulation demonstrates tangible efficiency gains. The framework offers a scalable, adaptable approach for ODMD and other city logistics services, with significant implications for responsive resource allocation and urban sustainability.

Abstract

Micro-delivery services offer promising solutions for on-demand city logistics, but their success relies on efficient real-time delivery operations and fleet management. On-demand meal delivery platforms seek to optimize real-time operations based on anticipatory insights into citywide demand distributions. To address these needs, this study proposes a short-term predict-then-cluster framework for on-demand meal delivery services. The framework utilizes ensemble-learning methods for point and distributional forecasting with multivariate features, including lagged-dependent inputs to capture demand dynamics. We introduce Constrained K-Means Clustering (CKMC) and Contiguity Constrained Hierarchical Clustering with Iterative Constraint Enforcement (CCHC-ICE) to generate dynamic clusters based on predicted demand and geographical proximity, tailored to user-defined operational constraints. Evaluations of European and Taiwanese case studies demonstrate that the proposed methods outperform traditional time series approaches in both accuracy and computational efficiency. Clustering results demonstrate that the incorporation of distributional predictions effectively addresses demand uncertainties, improving the quality of operational insights. Additionally, a simulation study demonstrates the practical value of short-term demand predictions for proactive strategies, such as idle fleet rebalancing, significantly enhancing delivery efficiency. By addressing demand uncertainties and operational constraints, our predict-then-cluster framework provides actionable insights for optimizing real-time operations. The approach is adaptable to other on-demand platform-based city logistics and passenger mobility services, promoting sustainable and efficient urban operations.

Figures (16)

• Figure 1: Workflow of the predict-then-cluster framework for supporting real-time operations in meal delivery services.
• Figure 2: The average number of orders received in the European city during different hours on weekdays versus weekends , from April $1^{st}$, 2020 to September $14^{th}$, 2020.
• Figure 3: The heatmap visualization of the number of orders received by different pick-up zones from the European use case per 15-minute intervals, averaged over three time periods: whole day, during lunch times, and dinner times. The heatmaps use a color scale to indicate the density of orders, with hotter colors representing higher zonal order volumes.
• Figure 4: The average number of orders received in Taiwan during different hours on weekdays versus weekends, was calculated using a dataset gathered over a three-month period. Error bars represent the standard deviation per hour across different days.
• Figure 5: The heatmap visualization of the number of orders received by different pick-up zones from the Taiwanese use case per 15-minute intervals, averaged over three time periods: whole day, during lunch times, and dinner times. The heatmaps use a color scale to indicate the density of orders, with hotter colors representing higher order volumes.