Mr. Qi Zhan is an emerging researcher whose work focuses on advancing intelligent, adaptive, and resource-efficient solutions across network architecture, network intelligence, and machine learning. His research addresses the increasing complexity of distributed packet-measurement systems within Computing and Network Convergence (CNC), where dynamic traffic patterns and heterogeneous resource conditions require optimized orchestration strategies. Mr. Zhan has developed a reinforcement learning–based orchestration method for distributed sketch deployment, offering a significant advancement over traditional static or heuristic allocation approaches. His method enables an intelligent agent to continuously sense node-level resource states and automatically adjust deployment decisions to achieve global optimization. Through this approach, he successfully reduces maximum resource utilization, minimizes the standard deviation of resource consumption, and achieves balanced packet-measurement load distribution across distributed nodes. These improvements enhance fairness, increase scalability, and promote more stable operational performance within large-scale programmable networks. His recent publication, Enabling Resource-Aware Distributed Sketch Deployment with Reinforcement Learning, demonstrates how reinforcement learning can effectively support fine-grained measurement tasks while reducing overhead and adapting to real-time network changes. By integrating machine learning with network system design, Mr. Zhan’s work contributes to the development of intelligent, self-optimizing network infrastructures capable of supporting future high-performance, data-driven, and computation-intensive distributed applications.
Profiles: Orcid
Featured Publication
Zhan, Q., Dong, Y., Tian, L., Hu, Y., Xia, J., Zhu, Y., Wang, Z., Guo, X., & Li, H. (2025). Enabling resource-aware distributed sketch deployment with reinforcement learning. Conference paper presented at the ACM.