Collaborative route optimization for efficient and constraint-aware express delivery using deep reinforcement learning

Haoqiang Liu

doi:10.1016/j.neucom.2026.133451

Collaborative route optimization for efficient and constraint-aware express delivery using deep reinforcement learning

2026-03-25

0

PRE

AI

DOI：10.1016/j.neucom.2026.133451

Original

Original request for help

Save

Abstract

En 中文

• Proposes a demand- and time-window-constrained K-means clustering algorithm to capture spatial–temporal order patterns in dynamic logistics systems. • Formulates collaborative courier scheduling as a multi-depot vehicle routing problem with simultaneous pickup–delivery and time windows (MDVRPSPDTW). • Develops a deep reinforcement learning-based encoder–decoder framework integrating graph neural networks, attention mechanisms, and REINFORCE optimization for adaptive, constraint-aware routing. • Achieves significant improvements over state-of-the-art routing strategies, demonstrating robust adaptability to real-world last-mile delivery dynamics.

Keywords:

demand-constrained clustering

time-window optimization

multi-depot vehicle routing

deep reinforcement learning

collaborative courier scheduling

Collaborative route optimization for efficient and constraint-aware express delivery using deep reinforcement learning

Abstract

Journal

Organization