Teacher Changfu Xu from the School of Software and IoT Engineering Publishes Significant Research Outcomes in the CCF A-level Journal IEEE Transactions on Mobile Computing.
Recently, Xu Changfu, a young teacher from the School of Software and IoT Engineering, published aresearch papertitled "EnhancingQoE in Collaborative Edge Systems with Feedback DiffusionGenerativeScheduling" in the top international journal in the field of mobile computing,IEEE Transactions on Mobile Computing.The paper lists Dr. Xu Changfu as the first author, with collaborators including Professor Jia Weijia, Professor Wang Tian,Associate Professor Guo Jianxiong, Dr. Zeng Jiandian, and Dr. LiangYuzhu from Beijing Normal University, as well as Professor CaoJiannong from Hong Kong Poly technic University, Professor Dai Haipengfrom Nanjing University, and Dr. Zou Haodong from Anhui University.
IEEE Transactions on Mobile Computing is a leading journal in the field of computer networks,focusing on academic research in mobile computing. It is one of the authoritative journals in the field of mobile computing and is classified as an A-level international academic journal recommended by the China Computer Federation(CCF).
This paper presents FDEdge, anovel feedback diffusion generative scheduling method to enhance user experience in collaborative edge systems. The FDEdge first designs aninnovative feedback diffusion model, which utilizes historical action probability information during the denoising process. The feedback diffusion model is then integrated into deep reinforcement learning,forming an efficient and effective framework for task scheduling in collaborative edge systems. Additionally, a probabilistic derivation analysis is provided to ensure the rationality of FDEdge. Extensive experimental results show that the proposed FDEdge method reducesservice latency by 45.42% to 87.57% compared to state-of-the-artmethods and accelerates the training episode duration by 2.5 times.This research provides new ideas and methods for the field of edge computing, with significant theoretical and practical value.