Wen Chen
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Wen Chen has a total of 26 co-authors. Here are the top co-authors by number of shared publications:
Suchao Xiao
2 shared publicationsXueqin Jiang
2 shared publicationsDi Jiang
State Grid Wuxi Power Supply Company,Wuxi,China
1 shared publication- Jiawei Liu
School of Information Science and Technology, Donghua University, Shanghai 201620, China
1 shared publication - Jinming Yu
School of Information Science and Technology, Donghua University, Shanghai 201620, China
1 shared publication Jun Li
School of Mechanical and Electrical Engineering, Guangzhou University, Guangzhou, China
1 shared publication
Latest Publications
- Open Access
Dynamic Edge Loading Balancing with Edge Node Activity Prediction and Accelerating the Model Convergence
In mobile edge computing networks, achieving effective load balancing across edge server nodes is essential for minimizing task processing latency. However, the lack of a priori knowledge regarding the current load state of edge nodes for user devices presents a significant challenge in multi-user, multi-edge node scenarios. This challenge is exacerbated by the inherent dynamics and uncertainty of edge node load variations. To tackle these issues, we propose a deep reinforcement learning-based approach for task offloading and resource allocation, aiming to balance the load on edge nodes while reducing the long-term average cost. Specifically, we decompose the optimization problem into two subproblems, task offloading and resource allocation. The Karush–Kuhn–Tucker (KKT) conditions are employed to derive the optimal strategy for communication bandwidth and computational resource allocation for edge nodes. We utilize Long Short-Term Memory (LSTM) networks to forecast the real-time activity of edge nodes. Additionally, we integrate deep compression techniques to expedite model convergence, facilitating faster execution on user devices. Our simulation results demonstrate that our proposed scheme achieves a 47% reduction in terms of the task drop rate, a 14% decrease in the total system cost, and a 7.6% improvement in the runtime compared to the baseline schemes.
Open AccessDynamic Edge Loading Balancing with Edge Node Activity Prediction and Accelerating the Model Convergence
In mobile edge computing networks, achieving effective load balancing across edge server nodes is essential for minimizing task processing latency. However, the lack of a priori knowledge regarding the current load state of edge nodes for user devices presents a significant challenge in multi-user, multi-edge node scenarios. This challenge is exacerbated by the inherent dynamics and uncertainty of edge node load variations. To tackle these issues, we propose a deep reinforcement learning-based approach for task offloading and resource allocation, aiming to balance the load on edge nodes while reducing the long-term average cost. Specifically, we decompose the optimization problem into two subproblems, task offloading and resource allocation. The Karush–Kuhn–Tucker (KKT) conditions are employed to derive the optimal strategy for communication bandwidth and computational resource allocation for edge nodes. We utilize Long Short-Term Memory (LSTM) networks to forecast the real-time activity of edge nodes. Additionally, we integrate deep compression techniques to expedite model convergence, facilitating faster execution on user devices. Our simulation results demonstrate that our proposed scheme achieves a 47% reduction in terms of the task drop rate, a 14% decrease in the total system cost, and a 7.6% improvement in the runtime compared to the baseline schemes.
- 13 June 2024
- Published by Elsevier in Computer Networks
- Vol. 250
Optimizing resource allocation for cluster D2D-assisted fog computing networks: A three-layer Stackelberg game approach
- 13 June 2024
- Published by Elsevier in Computer Networks
- Vol. 250
Optimizing resource allocation for cluster D2D-assisted fog computing networks: A three-layer Stackelberg game approach
- Open Access
- 26 May 2023
- Published by Association for Computing Machinery (ACM)
- p. 140-144
Join active and passive beamforming optimization for double-RIS aided CR networks
Intelligent Reflector Surface (IRS), also known as Reconfigurable Intelligent Surface (RIS), is a promising method to improve the energy efficiency of wireless communication system by intelligently reconfiguring the channel environment in recent years. Cognitive radio (CR) is an effective solution to improve the spectrum efficiency of wireless communication, which allows secondary users (SU) to share the spectrum with primary users (PU). In this paper, double-RIS are introduced into the downlink multi-input single-output (MISO) CR system to assist communication. Our design goal is to maximize the sum of the achievable rates of SU and PUs by jointly optimizing the beamforming of cognitive base station and main base station and the reflection coefficients of the double-RIS under the constraints of base station transmission power and interference threshold. In order to achieve this goal, we use semi-positive definite relaxation method (SDR) and alternate optimization method to get the optimal solutions of beamforming vector and reflection coefficient. The simulation results show that in the double-RIS aided CR networks, the proposed scheme can significantly improve the reachable rate of all users.
- 26 May 2023
- Published by Association for Computing Machinery (ACM)
- p. 140-144
Open AccessJoin active and passive beamforming optimization for double-RIS aided CR networks
Intelligent Reflector Surface (IRS), also known as Reconfigurable Intelligent Surface (RIS), is a promising method to improve the energy efficiency of wireless communication system by intelligently reconfiguring the channel environment in recent years. Cognitive radio (CR) is an effective solution to improve the spectrum efficiency of wireless communication, which allows secondary users (SU) to share the spectrum with primary users (PU). In this paper, double-RIS are introduced into the downlink multi-input single-output (MISO) CR system to assist communication. Our design goal is to maximize the sum of the achievable rates of SU and PUs by jointly optimizing the beamforming of cognitive base station and main base station and the reflection coefficients of the double-RIS under the constraints of base station transmission power and interference threshold. In order to achieve this goal, we use semi-positive definite relaxation method (SDR) and alternate optimization method to get the optimal solutions of beamforming vector and reflection coefficient. The simulation results show that in the double-RIS aided CR networks, the proposed scheme can significantly improve the reachable rate of all users.
- 1 April 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 2253-2257
Dielectric Loss of Natural Ester-paper Insulation Withstanding Voltage Transients
Natural ester-paper (NEP) insulation is a promising alternative to conventional mineral oil impregnated paper insulation, and its dielectric properties under transient voltages are of critical concern for future extensive application. Therefore, the dielectric loss of NEP insulation under pulse transient voltage is calculated in this paper. First, the frequency domain spectroscopy of the NEP insulation is used to establish its equivalent extended Debye circuit model. Then, under the impulse voltage excitation expressed by double exponential function, the voltage and current response of the model is solved to simulate the actual dielectric response of the paper, and then the dielectric response of NEP insulation is obtained. Finally, the influence of temperature and waveform characteristics on the power spectrum of the dielectric loss is analyzed.
- 1 April 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 2253-2257
Dielectric Loss of Natural Ester-paper Insulation Withstanding Voltage Transients
Natural ester-paper (NEP) insulation is a promising alternative to conventional mineral oil impregnated paper insulation, and its dielectric properties under transient voltages are of critical concern for future extensive application. Therefore, the dielectric loss of NEP insulation under pulse transient voltage is calculated in this paper. First, the frequency domain spectroscopy of the NEP insulation is used to establish its equivalent extended Debye circuit model. Then, under the impulse voltage excitation expressed by double exponential function, the voltage and current response of the model is solved to simulate the actual dielectric response of the paper, and then the dielectric response of NEP insulation is obtained. Finally, the influence of temperature and waveform characteristics on the power spectrum of the dielectric loss is analyzed.
- Open Access
Enhancing Mobile Edge Computing with Efficient Load Balancing Using Load Estimation in Ultra-Dense Network
With the exponential growth of mobile devices and the emergence of computationally intensive and delay-sensitive tasks, the enormous demand for data and computing resources has become a big challenge. Fortunately, the combination of mobile edge computing (MEC) and ultra-dense network (UDN) is considered to be an effective way to solve these challenges. Due to the highly dynamic mobility of mobile devices and the randomness of the work requests, the load imbalance between MEC servers will affect the performance of the entire network. In this paper, the software defined network (SDN) is applied to the task allocation in the MEC scenario of UDN, which is based on routing of corresponding information between MEC servers. Secondly, a new load balancing algorithm based on load estimation by user load prediction is proposed to solve the NP-hard problem in task offloading. Furthermore, a genetic algorithm (GA) is used to prove the effectiveness and rapidity of the algorithm. At present, if the load balancing algorithm only depends on the actual load of each MEC, it usually leads to ping-pong effect. It is worth mentioning that our method can effectively reduce the impact of ping-pong effect. In addition, this paper also discusses the subtask offloading problem of divisible tasks and the corresponding solutions. At last, simulation results demonstrate the efficiency of our method in balancing load among MEC servers and its ability to optimize systematic stability.
Open AccessEnhancing Mobile Edge Computing with Efficient Load Balancing Using Load Estimation in Ultra-Dense Network
With the exponential growth of mobile devices and the emergence of computationally intensive and delay-sensitive tasks, the enormous demand for data and computing resources has become a big challenge. Fortunately, the combination of mobile edge computing (MEC) and ultra-dense network (UDN) is considered to be an effective way to solve these challenges. Due to the highly dynamic mobility of mobile devices and the randomness of the work requests, the load imbalance between MEC servers will affect the performance of the entire network. In this paper, the software defined network (SDN) is applied to the task allocation in the MEC scenario of UDN, which is based on routing of corresponding information between MEC servers. Secondly, a new load balancing algorithm based on load estimation by user load prediction is proposed to solve the NP-hard problem in task offloading. Furthermore, a genetic algorithm (GA) is used to prove the effectiveness and rapidity of the algorithm. At present, if the load balancing algorithm only depends on the actual load of each MEC, it usually leads to ping-pong effect. It is worth mentioning that our method can effectively reduce the impact of ping-pong effect. In addition, this paper also discusses the subtask offloading problem of divisible tasks and the corresponding solutions. At last, simulation results demonstrate the efficiency of our method in balancing load among MEC servers and its ability to optimize systematic stability.
- 14 April 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1144-1147
Research on Fault Diagnosis Based on Virtual Prototype Technology of Mechanical System
The increasing complexity of the mechanical system makes it difficult for the corresponding physical prototype to diagnose and distinguish the faults. Therefore, the simulation research of virtual prototype is needed before the physical prototype experiment. Virtual prototype technology is a technology that uses computer technology, information technology and other technologies to carry out mathematical modeling of the corresponding mechanical system physical machine. This paper will take a mechanical system gearbox as the modeling object based on Virtual Prototype Technology for modeling analysis. Through the mathematical modeling of the gearbox, the typical fault features of the gearbox are extracted. Through the simulation of the actual working process of the gearbox, the corresponding load and constraints are imposed on the gearbox, and the corresponding collision parameters are selected to realize the dynamic simulation of the gearbox and observe its corresponding working conditions. Finally, the corresponding simulation results are applied to the actual work. The simulation results show that the virtual prototype technology is conducive to the early diagnosis and prevention of mechanical system faults, and has good industrial application value.
- 14 April 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1144-1147
Research on Fault Diagnosis Based on Virtual Prototype Technology of Mechanical System
The increasing complexity of the mechanical system makes it difficult for the corresponding physical prototype to diagnose and distinguish the faults. Therefore, the simulation research of virtual prototype is needed before the physical prototype experiment. Virtual prototype technology is a technology that uses computer technology, information technology and other technologies to carry out mathematical modeling of the corresponding mechanical system physical machine. This paper will take a mechanical system gearbox as the modeling object based on Virtual Prototype Technology for modeling analysis. Through the mathematical modeling of the gearbox, the typical fault features of the gearbox are extracted. Through the simulation of the actual working process of the gearbox, the corresponding load and constraints are imposed on the gearbox, and the corresponding collision parameters are selected to realize the dynamic simulation of the gearbox and observe its corresponding working conditions. Finally, the corresponding simulation results are applied to the actual work. The simulation results show that the virtual prototype technology is conducive to the early diagnosis and prevention of mechanical system faults, and has good industrial application value.
- 1 January 2020
- Published by Elsevier in Computers and Electrical Engineering
- Vol. 81
A DDoS attacks traceback scheme for SDN-based smart city
- 1 January 2020
- Published by Elsevier in Computers and Electrical Engineering
- Vol. 81
A DDoS attacks traceback scheme for SDN-based smart city
- 1 November 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 735-739
Dynamic Allocation of 5G Transport Network Slice Bandwidth Based on LSTM Traffic Prediction
5G networks will be characterized by the extremely wide bandwidth that will be available to the user. A more flexible transmission network is necessary to support the demand generated by the access network to continuously increase network bandwidth. Transport network slicing will be a promising technology to address those challenges. In this paper, we focused on the dynamic resource allocation problem of bandwidth in transport network slices. We introduce a novel LSTM -based traffic-predict dynamic transport network slicing framework (LSTM-TPDTNS). Our approach consists of two phases: the traffic prediction phase and the bandwidth configuration phase. For the first phase, we use long and short memory models to predict traffic. For the second phase, we model our problem as a fractional knapsack problem, and we use greedy algorithms to find approximate solutions. Dynamic allocation of resources to services of different priorities can be realized, thereby improving the service quality and user experience of the entire system.
- 1 November 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 735-739
Dynamic Allocation of 5G Transport Network Slice Bandwidth Based on LSTM Traffic Prediction
5G networks will be characterized by the extremely wide bandwidth that will be available to the user. A more flexible transmission network is necessary to support the demand generated by the access network to continuously increase network bandwidth. Transport network slicing will be a promising technology to address those challenges. In this paper, we focused on the dynamic resource allocation problem of bandwidth in transport network slices. We introduce a novel LSTM -based traffic-predict dynamic transport network slicing framework (LSTM-TPDTNS). Our approach consists of two phases: the traffic prediction phase and the bandwidth configuration phase. For the first phase, we use long and short memory models to predict traffic. For the second phase, we model our problem as a fractional knapsack problem, and we use greedy algorithms to find approximate solutions. Dynamic allocation of resources to services of different priorities can be realized, thereby improving the service quality and user experience of the entire system.
- 7 December 2017
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Wireless Communications Letters
- Vol. 7 (3) , 408-411
Two-Layer LDPC Codes for Low Complexity ML Detection in GSM Systems
Although Maximum-Likelihood (ML) detection is able to achieve optimal performance in Generalized Spatial Modulation (GSM) systems, its exhaustive search leads to an intractable computational complexity. In this letter, Two-Layer Low-Density Parity-Check (TL-LDPC) codes are proposed to reduce the combinations of Transmission Vectors (TV) in LDPC coded GSM systems. Simulation results show that the complexity of the ML detection in the TL-LDPC coded GSM systems is much lower than that of the ML detection in conventional LDPC coded GSM systems with a negligible performance loss.
- 7 December 2017
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Wireless Communications Letters
- Vol. 7 (3) , 408-411
Two-Layer LDPC Codes for Low Complexity ML Detection in GSM Systems
Although Maximum-Likelihood (ML) detection is able to achieve optimal performance in Generalized Spatial Modulation (GSM) systems, its exhaustive search leads to an intractable computational complexity. In this letter, Two-Layer Low-Density Parity-Check (TL-LDPC) codes are proposed to reduce the combinations of Transmission Vectors (TV) in LDPC coded GSM systems. Simulation results show that the complexity of the ML detection in the TL-LDPC coded GSM systems is much lower than that of the ML detection in conventional LDPC coded GSM systems with a negligible performance loss.
- 24 January 2017
- Published by Springer Nature in Multimedia Tools and Applications
- Vol. 76 (19) , 20049-20065
Fuzzy multiple attribute decision access scheme in heterogeneous wireless network
- 24 January 2017
- Published by Springer Nature in Multimedia Tools and Applications
- Vol. 76 (19) , 20049-20065
Fuzzy multiple attribute decision access scheme in heterogeneous wireless network