Selected work

Representative Papers

A Context-Free Smart Grid Model Using Complex System Approach

Oct 01, 20142014 IEEE/ACM 18th International Symposium on Distributed Simulation and Real Time Applications

Addressing the challenges of global optimization and lack of modeling generality in smart grids, this paper proposes a context-free complex-system modeling paradigm. It pioneers the cross-scale integration of game theory with classical optimization techniques—including linear and dynamic programming—to establish a multi-level collaborative optimization model spanning generation, transmission, and consumption. The approach simultaneously captures structural complexity and preserves objective invariance, thereby enhancing computational scalability and policy adaptability without compromising robustness. Through simulation-driven validation, the model demonstrates efficient, high-fidelity optimization performance across diverse grid scales and topologies. This work provides a scalable theoretical framework and simulation platform enabling low-cost, high-accuracy deployment of optimization strategies in real-world smart grids.

5 citationsRead paper

Hierarchical Clustering of Complex Energy Systems Using Pretopology

Nov 27, 2025SMARTGREENS/VEHITS

Modeling and automatically classifying energy consumption patterns across large-scale distributed regional buildings remains challenging due to spatial heterogeneity, temporal dynamics, and limited physical interpretability. Method: This paper proposes a multi-criteria hierarchical clustering method grounded in pre-topological space properties. It jointly models energy consumption profiles using spatial, temporal, and statistical features; employs pre-topological structures to encode shape and connectivity; and integrates Pearson correlation analysis for multi-dimensional feature fusion. An open-source Python library is developed to support point-set, time-series, and real-world energy data processing. Contribution/Results: The work pioneers the application of pre-topology to energy pattern recognition, enhancing both clustering accuracy and physical interpretability. Experiments demonstrate exact spatial cluster recovery on synthetic data, perfect time-series clustering (adjusted Rand index = 1.0), and robust grouping of 400 real-world energy sites—enabling a new paradigm for intelligent, low-intervention energy management.

2 citationsRead paper

Strong aggregation of the Markov chains associated with matching models based on the automorphism group of their compatibility graphs

Sep 15, 2026

研究通过图的自同构群分析匹配模型中的马尔可夫链强聚合性，扩展了贪婪和非贪婪匹配规则下的强聚合条件。

0 citationsRead paper

Sender and Receiver Energy Consumption in a Sensor Network

Jul 23, 2026

This study addresses the challenges of queue synchronization and energy management in sensor networks where both transmitters and receivers are subject to energy constraints. The authors develop a joint queueing model that integrates energy packets and data packets, wherein data transmission consumes discrete units of energy, and insufficient energy leads to transmission delays or reception failures. Innovatively, under bilateral energy constraints, they formulate a queueing network model admitting a product-form stationary solution. By leveraging Markov chain theory and queueing analysis, they establish sufficient conditions for system ergodicity. Furthermore, they propose a convergent and correct numerical algorithm that efficiently computes the stationary distribution and key traffic performance metrics.

0 citationsRead paper

Parallel gradient boosting for flexible estimation of conditional distributions

Jul 15, 2026

This work addresses the computational inefficiency of traditional gradient boosting in multi-output prediction tasks—such as multiple quantile regression—where a separate base model must be trained for each output. The authors propose a general and efficient parallel gradient boosting algorithm that shares a unified descent direction across all outputs, requiring only a single base model per iteration and thereby substantially reducing computational overhead. Their approach overcomes existing limitations on loss functions and base learner types, supporting arbitrary combinations, and establishes the first scalable framework for multi-output conditional distribution estimation. Experiments demonstrate that the method achieves predictive accuracy comparable to XGBoost while accelerating training by several orders of magnitude, and it outperforms current nonparametric and semiparametric methods in high-dimensional settings with mixed or missing covariates.

0 citationsRead paper

Recent publications

Latest Papers

Strong aggregation of the Markov chains associated with matching models based on the automorphism group of their compatibility graphs

Sep 15, 2026

研究通过图的自同构群分析匹配模型中的马尔可夫链强聚合性，扩展了贪婪和非贪婪匹配规则下的强聚合条件。

0 citationsRead paper

Sender and Receiver Energy Consumption in a Sensor Network

Jul 23, 2026

This study addresses the challenges of queue synchronization and energy management in sensor networks where both transmitters and receivers are subject to energy constraints. The authors develop a joint queueing model that integrates energy packets and data packets, wherein data transmission consumes discrete units of energy, and insufficient energy leads to transmission delays or reception failures. Innovatively, under bilateral energy constraints, they formulate a queueing network model admitting a product-form stationary solution. By leveraging Markov chain theory and queueing analysis, they establish sufficient conditions for system ergodicity. Furthermore, they propose a convergent and correct numerical algorithm that efficiently computes the stationary distribution and key traffic performance metrics.

0 citationsRead paper

Parallel gradient boosting for flexible estimation of conditional distributions

Jul 15, 2026

This work addresses the computational inefficiency of traditional gradient boosting in multi-output prediction tasks—such as multiple quantile regression—where a separate base model must be trained for each output. The authors propose a general and efficient parallel gradient boosting algorithm that shares a unified descent direction across all outputs, requiring only a single base model per iteration and thereby substantially reducing computational overhead. Their approach overcomes existing limitations on loss functions and base learner types, supporting arbitrary combinations, and establishes the first scalable framework for multi-output conditional distribution estimation. Experiments demonstrate that the method achieves predictive accuracy comparable to XGBoost while accelerating training by several orders of magnitude, and it outperforms current nonparametric and semiparametric methods in high-dimensional settings with mixed or missing covariates.

0 citationsRead paper

Benchmarking Large Language Models on Floating-Point Error Classification

Jun 30, 2026

This study addresses the limited capability of large language models (LLMs) in statically detecting floating-point errors in code by introducing InterFLOPBench, the first specialized benchmark comprising 90 C-language kernels and 1,130 test cases. The task is formulated as a multi-label classification problem encompassing six canonical categories of floating-point errors. The authors propose an evaluation framework based on multi-label F1-score and conduct a systematic assessment across 14 prominent LLMs. Results show that state-of-the-art models, such as Qwen3-32B and Gemini 2.5 Flash, achieve overall F1-scores exceeding 0.88, with explicit errors like division-by-zero detected at a rate of 0.8479. However, implicit errors—including underflow and cancellation—remain challenging, with recognition rates around 0.60–0.62, highlighting current limitations in LLMs’ understanding of complex numerical semantics.

0 citationsRead paper

Mixed-Precision For Energy Efficient Computations

Jun 26, 2026

This study addresses the substantial time and energy costs associated with high-precision scientific computing. The authors propose a strategy that integrates mixed-precision computation with computer arithmetic optimizations to significantly improve energy efficiency while rigorously preserving numerical accuracy. Experimental evaluation on two representative scientific computing benchmarks—ReactorSimulator and LULESH—demonstrates the effectiveness of the approach: on ReactorSimulator, both execution time and energy consumption are reduced by 30%, while on LULESH, execution time decreases by 30% and energy usage drops by 25%. These results validate the method’s practicality and efficacy for enabling high-fidelity, energy-efficient simulations in scientific computing.

0 citationsRead paper