Youthful Teacher Changfu Xu from the School of Software and IoT Engineering Publishes Significant Research Outcomes in the CCF A-level Journal IEEE

Recently, Xu Changfu,a youthful teacherfrom the School of Software and Internet of Things Engineering at Jiangxi University of Finance and Economics published a paper as the first author in the top international journal in the field of service computing,IEEE Transactions on Services Computing.The title of the paper is“Enhancing AIGC Service Efficiency with Adaptive Multi-Edge Collaboration in ADistributed System.”Paper link: https://IEEE Xplore.ieee.org/document/11271533.

This paper innovativelypresentsan adaptive multi-edge collaboration method to improve the service efficiency of generative AI applications in distributed systems. Specifically, it first models the adaptive multi-edge server collaboration AIGC task offloading problem in a heterogeneous edge environment and solves it by dividing the problem into two stages: edge server selection and workload allocation. In the first stage, based on the traditional deep reinforcement learning DQN model, an adaptive multi-edge service selection model is constructed to obtain the indices of all available edge servers in the system. In the second stage, based on the first stage, aClosed-formWorkloadAllocationalgorithm (CWA) is designed to achieve the optimal allocation of workloads among all selected edge servers. Based on this approach, anAdaptiveMulti-EdgeCollaboration(AMCoEdge)method is formed, which can fully utilize the computing power of all edge servers to collaboratively accelerate the processing of generative AI tasks, minimizing task completion latency. Theoretical analysis shows thatthe proposed method can achieve near-optimal scheduling decisions with approximately linear time complexity. In addition, a more efficientHighCWA(HCWA) algorithmis designed,formingto an even more stable and efficient adaptive multi-edge collaboration methodnamedAMCoEdge-H.Finally, a distributed edge prototype system is developed to implement and validate the effectiveness of the proposed methods.

Journal Introduction：IEEE Transactions on Services Computing is a top journal in the fields of software engineering/system software/programming languages, focusing on academic research in the field of service computing. It is one ofthe authoritative journals in the field of service computing and is classified as an A-level international academic journal recommended by CCF (China Computer Federation).