GNN-PO: A Graph Neural Network Based Intelligent Task Allocation Model for Cloud Computing Using Puma Optimization

Mohamed Loey

doi:10.1016/j.suscom.2026.101481

GNN-PO: A Graph Neural Network Based Intelligent Task Allocation Model for Cloud Computing Using Puma Optimization

2026-09-17

0

PRE

AI

DOI：10.1016/j.suscom.2026.101481

Original

Original request for help

Save

Abstract

En 中文

• Proposes a GNN-PO framework for intelligent task allocation in cloud computing. • Graph neural networks learn task-resource relationships in heterogeneous clouds. • Puma Optimization minimizes makespan, waiting time, and SLA violations. • Evaluated on Alibaba Cluster Trace v2023 with 50,000 real-world tasks. • Outperforms GA and PSO in convergence, scalability, and resource utilization.

Keywords:

Graph Neural Network

Cloud Task Scheduling

Puma Optimization

Resource Allocation

SLA Management

GNN-PO: A Graph Neural Network Based Intelligent Task Allocation Model for Cloud Computing Using Puma Optimization

Abstract

Journal

Organization

Cited Papers