Design and Performance Evaluation of a Hybrid Task Scheduling Strategy for E-Commerce Logistics in Cloud-Edge Collaborative Environments

INTRODUCTION: The rapid growth of e-commerce has led to highly concurrent and diverse logistics operations, posing significant challenges for task scheduling under dynamic network conditions and heterogeneous computing resources. Conventional cloud-centric scheduling lacks real-time responsiveness, while purely edge-based decisions suffer from limited global visibility, resulting in suboptimal performance. OBJECTIVES: This study addresses these limitations by designing a cloud-edge collaborative hybrid scheduling method tailored for e-commerce logistics, aiming to simultaneously minimize latency, reduce energy consumption, and maximize task completion rates under fluctuating workloads. METHODS: The proposed framework integrates three core components: task encoding to capture heterogeneity, node state prediction using lightweight temporal models, and multi-objective scheduling driven by reinforcement learning. This enables the system to adapt dynamically to changes in bandwidth, node load, and task urgency. RESULTS: Evaluated under simulated peak and off-peak e-commerce scenarios, the method outperforms baseline approaches by reducing average task latency by 18.7%, increasing completion rate by 9.4%, and cutting system-wide energy consumption by 12.3%. CONCLUSION: By effectively coordinating cloud and edge resources, the approach provides a robust foundation for building low-latency, energy-efficient, and reliable scheduling systems, with practical implications for warehouse automation, instant delivery networks, and other time-sensitive logistics applications.