jingli zhou

jingli zhou

orcid

Research Focus & Metrics

Academic Career Length

2Years (2025-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
20 February 2025 in IEEE Transactions on Cognitive Communications and Networking
Verified
Journal Article
Closed access
YSYuqian Song
Yuqian Song
JZjingli zhou
jingli zhou
XLXinyuan Li
Xinyuan Li
JLJun Liu
Jun Liu
Packet routing in modern communication networks faces increasing challenges due to evolving network dynamics and stringent quality-of-service (QoS) requirements. Traditional rule-based and model-based routing methods, constrained by pre-defined rules and static network assumptions, often fail to address these complexities. This paper presents an experience-driven, model-free routing algorithm grounded in deep reinforcement learning (DRL), offering adaptive and efficient forwarding decisions. For practical deployment, we incorporate GraphTransformer networks into the DRL decision framework to enhance generalization under diverse and unseen network conditions. Specifically, we explore two element-level mapping paradigms, node-based and edge-based, to identify the optimal configuration. In addition, we design a hybrid action space that effectively incorporates domain-specific knowledge, thereby contributing to a more robust DRL agent in dynamic routing environments. Simulation results indicate that our proposed algorithms demonstrate significant robustness and generalization capabilities, revealing key principles for designing an optimal GNN-enhanced DRL-based routing algorithm.

24 January 2025 in Electronics
Verified
Journal Article
Open access
PDF file
XLXinyuan Li
Xinyuan Li
JLJunze Li
Junze Li
JZjingli zhou
jingli zhou
JLJun Liu
Jun Liu
Deep Reinforcement Learning (DRL) has demonstrated promising capabilities for routing optimization in Software-Defined Networks (SDNs). However, existing DRL-based routing algorithms are struggling to extract graph-structured information and constrained to a fixed topology, suffering from the lack of robustness. In this paper, we strengthen the advantages of Graph Neural Networks (GNNs) for DRL-based routing optimization and propose a novel algorithm named Graph Transformer Star Routing (GTSR) to enhance robustness against topology changes. GTSR utilizes the multi-agent architecture to enable each node to make routing decisions independently, and introduces a Graph Transformer to equip agents with the capabilities of handling topology changes. Furthermore, we carefully design a global message-passing mechanism with a virtual star node and a path-based readout method, enhancing the long-range perception of traffic and the detection of potential congestion for routing decision-making. Moreover, we construct a multi-agent cooperation mechanism to facilitate the learning of universal perceptual strategies and reduce the amount of computation. Extensive experiments on multiple real-world network topologies demonstrate that GTSR is capable of adapting to unseen topology changes without retraining and decreases end-to-end latency by at least 47% and packet loss rate by at least 10% compared to all baselines, highlighting strong robustness.

8 September 2025 in IEEE Transactions on Mobile Computing
Verified
Journal Article
Closed access
YSYuqian Song
Yuqian Song
JZjingli zhou
jingli zhou
SYShudan Yu
Shudan Yu
JLJun Liu
Jun Liu
The Internet today hosts a multitude of communication sessions from diverse vertical industries, each with distinct and increasingly stringent quality-of-service (QoS) requirements across multiple performance metrics. However, QoS-aware routing remains a significant challenge in traffic engineering, as existing solutions struggle to adapt to dynamic network conditions and meet these rigorous QoS demands. To address this issue, this paper proposes a multi-agent graph reinforcement learning-based routing algorithm that provides differentiated treatment for multiple services. First, we explore both node-based and link-based graph reinforcement learning paradigms for performance comparison. Second, two key mechanisms, i.e., a packet sacrifice mechanism and a Tchebycheff-based reward function, are designed to realize adaptive sacrifice behavior patterns, aiming to optimize the lower bound of service satisfaction rates and enhance fairness across services. Furthermore, to ensure practical applicability, we devise a distributed computing architecture featuring neighborhood-restricted data acquisition and asynchronous historical information retrieval. Extensive simulation results demonstrate that our proposed algorithms significantly outperform benchmark methods regarding the minimum service satisfaction rate, even under unseen networks. Besides, the distributed computing architecture is proven to incur no performance penalty, which can be generalized to other resource-constrained applications.

10 December 2025 in IEEE Transactions on Cognitive Communications and Networking
Verified
Journal Article
Closed access
YSYuqian Song
Yuqian Song
JZjingli zhou
jingli zhou
SYShudan Yu
Shudan Yu
JLJun Liu
Jun Liu
Internet service providers (ISPs) require modern networks to deliver high service satisfaction across diverse traffic types while maintaining minimal operational overhead. Traditional routing algorithms, although computationally efficient, rely on fixed heuristics and often fail to accommodate increasingly dynamic and heterogeneous network demands. In contrast, deep reinforcement learning (DRL)-based routing algorithms offer greater flexibility but typically introduce considerable computational cost. This issue is especially pronounced in prevalent centralized software-defined networking (SDN) architectures, resulting in excessive overhead and increased latency as the network grows. In this paper, we model a hybrid-traffic network scenario and propose a quality of service (QoS)-aware and low-cost routing algorithm based on graph reinforcement learning within a hybrid knowledge-defined networking (KDN) architecture. Routing decisions are jointly made by a global controller and distributed forwarding devices, with the unified objective of maximizing the minimum service satisfaction rate across the network. To reduce computational overhead and minimize disruptions from frequent route updates, we investigate two efficient implementation strategies: partial deployment (horizontal) and intermittent activation (vertical). Experimental results demonstrate that the proposed algorithm significantly outperforms benchmarks regarding the minimum service satisfaction rate, even under previously unseen topologies. Remarkably, our evaluation shows that 50% of the maximum achievable performance gain can be obtained through a lightweight implementation where just 28.5% of forwarding devices conduct policy inference, operating for only 60% of the total run time.

12 November 2025 in IEEE Internet of Things Journal
Verified
Journal Article
Closed access
YSYuqian Song
Yuqian Song
JZjingli zhou
jingli zhou
SYShudan Yu
Shudan Yu
JLJun Liu
Jun Liu
The rapid proliferation of end devices and their hosted applications has significantly increased the demand for data processing. However, lightweight edge endpoints often suffer from limited computational capabilities due to power and battery constraints. Consequently, computation-intensive tasks are typically offloaded to cloud servers located in data centers for processing. While existing studies often abstract the edge–cloud connection as a direct link, the actual backbone network features complex topologies and constrained bandwidth, which intensify as workload scales up. To address this challenge, this paper investigates routing optimization within the backbone network under task offloading scenarios, and proposes a workload-aware deep reinforcement learning (DRL)-based routing algorithm for joint minimization of transmission and processing delays. Specifically, we first develop a detailed system model using discrete event simulation. Then, we design an integrated DRL-based decision-making scheme that simultaneously determines optimal routing paths and target data centers. Extensive experiments on diverse real-world network topologies and heterogeneous task workloads demonstrate that the proposed algorithm significantly outperforms conventional baselines in reducing both transmission and processing delays.