Chenye Wu
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
Chenye Wu has a total of 177 co-authors. Here are the top co-authors by number of shared publications:
- Chenbei Lu
Institute for Interdisciplinary Information Sciences, Tsinghua University, Beijing, China
17 shared publications - Dan Wang
Hong Kong University of Science and Technology, Kowloon, Hong Kong
9 shared publications - Jian Shi
Department of Engineering Technology, University of Houston, Houston, TX, USA
9 shared publications - Nan Gu
Institute for Interdisciplinary Information Sciences, Tsinghua University, Beijing, China
9 shared publications - Jingshi Cui
Institute for Interdisciplinary Information Sciences, Tsinghua University, Beijing, China
7 shared publications Jianwei Huang
Department of Information Engineering, The Chinese University of Hong Kong, Shatin, New Territories, Hong Kong
6 shared publications
Latest Publications
- 1 November 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Smart Grid
- Vol. 17  (6) , 5983-5998
Edge AI-Enabled PMU Event Classification With Knowledge Distillation and Federated Learning
- 1 November 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Smart Grid
- Vol. 17  (6) , 5983-5998
Edge AI-Enabled PMU Event Classification With Knowledge Distillation and Federated Learning
- 2 July 2026
- Published by Elsevier in Applied Energy
- Vol. 422
Manifold optimized GAN based EV charging session data generation
- 2 July 2026
- Published by Elsevier in Applied Energy
- Vol. 422
Manifold optimized GAN based EV charging session data generation
- 30 June 2026
- Published by Elsevier in Electric Power Systems Research
- Vol. 262
Deciphering convex hull pricing: effectively aligning demand with electricity prices enabled by approximated minkowski sum
- 30 June 2026
- Published by Elsevier in Electric Power Systems Research
- Vol. 262
Deciphering convex hull pricing: effectively aligning demand with electricity prices enabled by approximated minkowski sum
- 25 May 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Industrial Informatics
- Vol. PP  (15513203) , 1-12
TACOS: A Topology-Aware Cooperative Multiagent Framework for Power-Computing Coscheduling in Data Center Microgrids
The rapid expansion of data center (DC) clusters has led to a significant increase in electricity consumption and carbon emissions, making power-computing coscheduling a critical enabler for sustainable DC operation. Large-scale AI services introduce massive and bursty workloads with complex task dependencies, which pose substantial challenges to minute-level scheduling in DC microgrids integrated with renewable energy and energy storage. These challenges are further aggravated by multidimensional uncertainties, including stochastic job arrivals, volatile electricity prices, and photovoltaic generation, as well as by stringent quality-of-service requirements and time-coupled computation-energy constraints. Existing analytical optimization and single-agent learning methods are often limited by poor scalability, weak topology awareness of job structures, and insufficient safety guarantees under temporally coupled constraints. To address these issues, this article proposes a topology-aware heterogeneous multiagent deep reinforcement learning framework for dynamic power-computing co-optimization in DC microgrids. Specifically, the problem is formulated as a constrained Markov Game to characterize interactions among scheduling and energy management entities. A Graphormer-based representation module is developed to encode directed acyclic graph job topologies and capture task dependencies. Moreover, constrained heterogeneous-agent proximal policy optimization with hybrid advantage decomposition and sequential updates is designed to handle temporal coupling constraints and improve the stability and safety of policy learning. Experimental results under realistic AI workload and energy scenarios demonstrate that the proposed method can effectively reduce energy costs while satisfying task deadline constraints, and exhibits strong scalability and robustness in large-scale DC microgrids.
- 25 May 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Industrial Informatics
- Vol. PP  (15513203) , 1-12
TACOS: A Topology-Aware Cooperative Multiagent Framework for Power-Computing Coscheduling in Data Center Microgrids
The rapid expansion of data center (DC) clusters has led to a significant increase in electricity consumption and carbon emissions, making power-computing coscheduling a critical enabler for sustainable DC operation. Large-scale AI services introduce massive and bursty workloads with complex task dependencies, which pose substantial challenges to minute-level scheduling in DC microgrids integrated with renewable energy and energy storage. These challenges are further aggravated by multidimensional uncertainties, including stochastic job arrivals, volatile electricity prices, and photovoltaic generation, as well as by stringent quality-of-service requirements and time-coupled computation-energy constraints. Existing analytical optimization and single-agent learning methods are often limited by poor scalability, weak topology awareness of job structures, and insufficient safety guarantees under temporally coupled constraints. To address these issues, this article proposes a topology-aware heterogeneous multiagent deep reinforcement learning framework for dynamic power-computing co-optimization in DC microgrids. Specifically, the problem is formulated as a constrained Markov Game to characterize interactions among scheduling and energy management entities. A Graphormer-based representation module is developed to encode directed acyclic graph job topologies and capture task dependencies. Moreover, constrained heterogeneous-agent proximal policy optimization with hybrid advantage decomposition and sequential updates is designed to handle temporal coupling constraints and improve the stability and safety of policy learning. Experimental results under realistic AI workload and energy scenarios demonstrate that the proposed method can effectively reduce energy costs while satisfying task deadline constraints, and exhibits strong scalability and robustness in large-scale DC microgrids.
- Open Access
- 12 May 2026
- Published by Elsevier in SSRN Electronic Journal
Mechanism-Guided Graph Learning for Carbon Border Policy Evaluation in Interconnected European Electricity Markets
The European Union's Carbon Border Adjustment Mechanism (CBAM) has moved from reporting to financial implementation, but its effects on interconnected electricity markets remain difficult to assess at the hourly scale used for dispatch and cross-border trade. Existing equilibrium models offer long-run policy insights but usually operate at aggregate temporal resolution, while standard spatiotemporal learning models can predict market variables but do not explain which market mechanisms drive the predicted response. This paper studies how a CBAM-type carbon-cost signal affects hourly electricity prices and carbon intensity across interconnected European electricity markets, and which modeled transmission channels explain the projected response. We develop a Mechanism-Guided Spatiotemporal Graph Neural Network (MG-STGNN) for dual-target prediction of electricity price and carbon intensity. The model represents three economically motivated channels: merit-order dispatch adjustment, cross-border trade arbitrage, and direct carbon-cost pass-through. Using six years of hourly ENTSO-E data from fifteen European countries, MG-STGNN achieves the strongest performance among ten tested models, with a price RMSE of 24.47~EUR/MWh and a carbon-intensity RMSE of 21.63~kg~CO$_2$/MWh. Under the full-CBAM scenario, projected price impacts range from $+0.7$~EUR/MWh in Sweden to $+7.8$~EUR/MWh in Poland. The model attributes the largest response shares to merit-order effects in several high-carbon Continental systems, trade adjustment in low-carbon peripheral systems, and carbon-cost pass-through in France and Italy. These results support differentiated policy screening by linking each country’s projected exposure to domestic dispatch adjustment, cross-border trade reallocation, or direct carbon-cost transmission, while remaining conditional on the modeled counterfactual design.
- 12 May 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessMechanism-Guided Graph Learning for Carbon Border Policy Evaluation in Interconnected European Electricity Markets
The European Union's Carbon Border Adjustment Mechanism (CBAM) has moved from reporting to financial implementation, but its effects on interconnected electricity markets remain difficult to assess at the hourly scale used for dispatch and cross-border trade. Existing equilibrium models offer long-run policy insights but usually operate at aggregate temporal resolution, while standard spatiotemporal learning models can predict market variables but do not explain which market mechanisms drive the predicted response. This paper studies how a CBAM-type carbon-cost signal affects hourly electricity prices and carbon intensity across interconnected European electricity markets, and which modeled transmission channels explain the projected response. We develop a Mechanism-Guided Spatiotemporal Graph Neural Network (MG-STGNN) for dual-target prediction of electricity price and carbon intensity. The model represents three economically motivated channels: merit-order dispatch adjustment, cross-border trade arbitrage, and direct carbon-cost pass-through. Using six years of hourly ENTSO-E data from fifteen European countries, MG-STGNN achieves the strongest performance among ten tested models, with a price RMSE of 24.47~EUR/MWh and a carbon-intensity RMSE of 21.63~kg~CO$_2$/MWh. Under the full-CBAM scenario, projected price impacts range from $+0.7$~EUR/MWh in Sweden to $+7.8$~EUR/MWh in Poland. The model attributes the largest response shares to merit-order effects in several high-carbon Continental systems, trade adjustment in low-carbon peripheral systems, and carbon-cost pass-through in France and Italy. These results support differentiated policy screening by linking each country’s projected exposure to domestic dispatch adjustment, cross-border trade reallocation, or direct carbon-cost transmission, while remaining conditional on the modeled counterfactual design.
- 12 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Power Systems
- Vol. PP  (08858950) , 1-14
Dynamic Reserve Sizing with Statistically Feasible Guarantees: An Uncertainty-Adaptive Robust Optimization Framework
The integration of intermittent renewable energy and increasing load volatility necessitates adaptive reserve sizing frameworks to balance grid reliability and economic efficiency. Static reserve allocation methods struggle with dynamic uncertainty, risking high costs or compromised security. This study introduces a probabilistic reserve sizing framework with three innovations: 1) a dynamic architecture for economic dispatch and Unit Commitment (UC) employing a unified joint chance constraint to manage system-wide, temporally-coupled risks under uncertainty; 2) a dual-stage statistically feasible robust optimization (RO) approach employing a novel unified dimension-collapsing map, enabling tractable learning-calibrated uncertainty set construction with dimension-free out-of-sample reliability guarantees; and 3) a Learning-to-Optimize acceleration strategy that transforms the robust UC into a tractable linear program, enabling scalability. Validation on standard test systems up to 118 buses demonstrates the framework's ability to dynamically adjust reserves through uncertainty-adaptive quantification. The solution reconciles statistical feasibility with computational efficiency, achieving scalable, cost-effective, and reliable operations. By integrating statistically feasible RO with machine learning acceleration, this framework provides a practical approach to managing uncertainty and ensuring economic and operational balance in power systems with high renewable penetration.
- 12 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Power Systems
- Vol. PP  (08858950) , 1-14
Dynamic Reserve Sizing with Statistically Feasible Guarantees: An Uncertainty-Adaptive Robust Optimization Framework
The integration of intermittent renewable energy and increasing load volatility necessitates adaptive reserve sizing frameworks to balance grid reliability and economic efficiency. Static reserve allocation methods struggle with dynamic uncertainty, risking high costs or compromised security. This study introduces a probabilistic reserve sizing framework with three innovations: 1) a dynamic architecture for economic dispatch and Unit Commitment (UC) employing a unified joint chance constraint to manage system-wide, temporally-coupled risks under uncertainty; 2) a dual-stage statistically feasible robust optimization (RO) approach employing a novel unified dimension-collapsing map, enabling tractable learning-calibrated uncertainty set construction with dimension-free out-of-sample reliability guarantees; and 3) a Learning-to-Optimize acceleration strategy that transforms the robust UC into a tractable linear program, enabling scalability. Validation on standard test systems up to 118 buses demonstrates the framework's ability to dynamically adjust reserves through uncertainty-adaptive quantification. The solution reconciles statistical feasibility with computational efficiency, achieving scalable, cost-effective, and reliable operations. By integrating statistically feasible RO with machine learning acceleration, this framework provides a practical approach to managing uncertainty and ensuring economic and operational balance in power systems with high renewable penetration.
- 12 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Systems Journal
- Vol. 20  (1) , 63-74
Strategic Decision-Making Under Uncertainty Through Bilevel Game Theory and Distributionally Robust Optimization
In strategic scenarios where decision-makers operate at different hierarchical levels, traditional optimization methods often fail to adequately handle uncertainties arising from incomplete information or unpredictable external factors. To fill this gap, we introduce a mathematical framework that integrates bilevel game theory with distributionally robust optimization (DRO), particularly suited for complex network systems. Our approach leverages the hierarchical structure of bilevel games to model leader–follower interactions while incorporating distributional robustness to guard against worst-case probability distributions. To ensure computational tractability, the Karush–Kuhn–Tucker (KKT) conditions are used to transform the bilevel challenge into a more manageable single-level model, and the infinite-dimensional DRO problem is reformulated into a finite equivalent. We propose a generalized algorithm to solve this integrated model. Simulation results validate our framework’s efficacy, demonstrating that under high uncertainty, the proposed model achieves up to a 22% cost reduction compared to traditional stochastic methods while maintaining a service level of over 90%. This highlights its potential to significantly improve decision quality and robustness in networked systems such as transportation and communication networks.
- 12 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Systems Journal
- Vol. 20  (1) , 63-74
Strategic Decision-Making Under Uncertainty Through Bilevel Game Theory and Distributionally Robust Optimization
In strategic scenarios where decision-makers operate at different hierarchical levels, traditional optimization methods often fail to adequately handle uncertainties arising from incomplete information or unpredictable external factors. To fill this gap, we introduce a mathematical framework that integrates bilevel game theory with distributionally robust optimization (DRO), particularly suited for complex network systems. Our approach leverages the hierarchical structure of bilevel games to model leader–follower interactions while incorporating distributional robustness to guard against worst-case probability distributions. To ensure computational tractability, the Karush–Kuhn–Tucker (KKT) conditions are used to transform the bilevel challenge into a more manageable single-level model, and the infinite-dimensional DRO problem is reformulated into a finite equivalent. We propose a generalized algorithm to solve this integrated model. Simulation results validate our framework’s efficacy, demonstrating that under high uncertainty, the proposed model achieves up to a 22% cost reduction compared to traditional stochastic methods while maintaining a service level of over 90%. This highlights its potential to significantly improve decision quality and robustness in networked systems such as transportation and communication networks.
- 11 November 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Power Systems
- Vol. 41  (3) , 1649-1661
Multi-Period Adaptive Robust Scheduling of Flexibility Providers With Diverse Response Delays
Flexibility providers play a crucial role in balancing the rapidly changing supply and demand in power systems. However, coordinating these heterogeneous providers necessitates a thorough understanding of their diverse response delays, which complicates effective scheduling. To address these challenges, we propose a multi-period adaptive robust framework designed to optimize the scheduling of flexibility providers while accounting for their varied response times. Our framework incorporates updated demand information, allowing for timely adjustments to scheduling decisions and aiming to maximize the total revenue of the system operator. We employ a mixed-integer linear programming (MILP) approach, where decision variables represent different types of flexibility providers, categorized as either continuous or discrete. To balance computational efficiency, optimality, and scalability, we utilize a partition-and-bound method that iteratively refines partitions of the uncertain demand sets. Numerical studies validate the effectiveness of our algorithm and underscore the importance of characterizing heterogeneous response delays in adaptive scheduling.
- 11 November 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Power Systems
- Vol. 41  (3) , 1649-1661
Multi-Period Adaptive Robust Scheduling of Flexibility Providers With Diverse Response Delays
Flexibility providers play a crucial role in balancing the rapidly changing supply and demand in power systems. However, coordinating these heterogeneous providers necessitates a thorough understanding of their diverse response delays, which complicates effective scheduling. To address these challenges, we propose a multi-period adaptive robust framework designed to optimize the scheduling of flexibility providers while accounting for their varied response times. Our framework incorporates updated demand information, allowing for timely adjustments to scheduling decisions and aiming to maximize the total revenue of the system operator. We employ a mixed-integer linear programming (MILP) approach, where decision variables represent different types of flexibility providers, categorized as either continuous or discrete. To balance computational efficiency, optimality, and scalability, we utilize a partition-and-bound method that iteratively refines partitions of the uncertain demand sets. Numerical studies validate the effectiveness of our algorithm and underscore the importance of characterizing heterogeneous response delays in adaptive scheduling.