Source author record

Junhua Zhao

Junhua Zhao appears in the imported research catalog. Authorship, coauthor and topic links are available while profile ownership is still unclaimed.

ResearcherUnclaimed source record

cond-mat.mtrl-sci Artificial Intelligence Computational Engineering, Finance, and Science eess.SP Machine Learning Multiagent Systems

Catalog footprint

What is connected

7works

6topics

4close collaborators

Actions

Connect this record

Open graph Browse works

Inspect adjacent papers, topics, institutions and collaborators without losing the researcher page.

Building this map preview

BZPEER is loading the nearby papers, people, topics and institutions for this page.

preprint2026arXiv

EngiAgent: Fully Connected Coordination of LLM Agents for Solving Open-ended Engineering Problems with Feasible Solutions

Engineering problem solving is central to real-world decision-making, requiring mathematical formulations that not only represent complex problems but also produce feasible solutions under data and physical constraints. Unlike mathematical problem solving, which operates on predefined formulations, engineering tasks demand open-ended analysis, feasibility-driven modeling, and iterative refinement. Although large language models (LLMs) have shown strong capabilities in reasoning and code generation, they often fail to ensure feasibility, which limits their applicability to engineering problem solving. To address this challenge, we propose EngiAgent, a multi-agent system with a fully connected coordinator that simulates expert workflows through specialized agents for problem analysis, modeling, verification, solving, and solution evaluation. The fully connected coordinator enables flexible feedback routing, overcoming the rigidity of prior pipeline-based reflection methods and ensuring feasibility at every stage of the process. This design not only improves robustness to diverse failure cases such as data extraction errors, constraint inconsistencies, and solver failures, but also enhances the overall quality of problem solving. Empirical results across four representative domains demonstrate that EngiAgent achieves substantial improvements in feasibility compared to prior approaches, establishing a new paradigm for feasibility-oriented engineering problem solving with LLMs. Our source code and data are available at https://github.com/AI4Engi/EngiAgent.

preprint2024arXiv

Exploring Large Language Model based Intelligent Agents: Definitions, Methods, and Prospects

Intelligent agents stand out as a potential path toward artificial general intelligence (AGI). Thus, researchers have dedicated significant effort to diverse implementations for them. Benefiting from recent progress in large language models (LLMs), LLM-based agents that use universal natural language as an interface exhibit robust generalization capabilities across various applications -- from serving as autonomous general-purpose task assistants to applications in coding, social, and economic domains, LLM-based agents offer extensive exploration opportunities. This paper surveys current research to provide an in-depth overview of LLM-based intelligent agents within single-agent and multi-agent systems. It covers their definitions, research frameworks, and foundational components such as their composition, cognitive and planning methods, tool utilization, and responses to environmental feedback. We also delve into the mechanisms of deploying LLM-based agents in multi-agent systems, including multi-role collaboration, message passing, and strategies to alleviate communication issues between agents. The discussions also shed light on popular datasets and application scenarios. We conclude by envisioning prospects for LLM-based agents, considering the evolving landscape of AI and natural language processing.

preprint2021arXiv

Super-Resolution Perception for Industrial Sensor Data

In this paper, we present the problem formulation and methodology framework of Super-Resolution Perception (SRP) on industrial sensor data. Industrial intelligence relies on high-quality industrial sensor data for system control, diagnosis, fault detection, identification, and monitoring. However, the provision of high-quality data may be expensive in some cases. In this paper, we propose a novel machine learning problem -- the SRP problem as reconstructing high-quality data from unsatisfactory sensor data in industrial systems. Advanced generative models are then proposed to solve the SRP problem. This technology makes it possible to empower existing industrial facilities without upgrading existing sensors or deploying additional sensors. We first mathematically formulate the SRP problem under the Maximum a Posteriori (MAP) estimation framework. A case study is then presented, which performs SRP on smart meter data. A network, namely SRPNet, is proposed to generate high-frequency load data from low-frequency data. We further employ a novel recognition-based loss and relativistic adversarial loss to constraint the reconstruction of waveforms explicitly. Experiments demonstrate that our SRP model can reconstruct high-frequency data effectively. Moreover, the reconstructed high-frequency data can lead to better appliance monitoring results without changing the monitoring appliances.

preprint2016arXiv

Superhigh moduli and tension-induced phase transition of monolayer gamma-boron at finite temperatures

Two dimensional (2D) gamma-boron (γ-B28) thin films have been firstly reported by the experiments of the chemical vapor deposition in the latest study [Tai et al., Angew. Chem. Int. Ed. 54, 1-6 (2015)]. However, their mechanical properties are still not clear. Here we predict the superhigh moduli (1460 GPa at 1 K and 744 GPa at 300 K) and the tension-induced phase transition of monolayer γ-B28 along a zigzag direction for large deformations at finite temperatures using molecular dynamics (MD) simulations. The new phase can be kept stable after unloading process at these temperatures. The predicted mechanical properties are reasonable with our results from density functional theory. This study provides physical insights into the origins of the new phase transition of monolayer γ-B28 at finite temperatures.

preprint2015arXiv

Thermal conductivity of graphene kirigami: ultralow and strain robustness

Kirigami structure, from the macro- to the nanoscale, exhibits distinct and tunable properties from original 2-dimensional sheet by tailoring. In present work, the extreme reduction of the thermal conductivity by tailoring sizes in graphene nanoribbon kirigami (GNR-k) is demonstrated using nonequilibrium molecular dynamics simulations. The results show that the thermal conductivity of GNR-k (around 5.1 Wm-1K-1) is about two orders of magnitude lower than that of the pristine graphene nanoribbon (GNR) (around 151.6 Wm-1K-1), while the minimum value is expected to be approaching zero in extreme case from our theoretical model. To explore the origin of the reduction of the thermal conductivity, the micro-heat flux on each atoms of GNR-k has been further studied. The results attribute the reduction of the thermal conductivity to three main sources as: the elongation of real heat flux path, the overestimation of real heat flux area and the phonon scattering at the vacancy of the edge. Moreover, the strain engineering effect on the thermal conductivity of GNR-k and a thermal robustness property has been investigated. Our results provide physical insights into the origins of the ultralow and robust thermal conductivity of GNR-k, which also suggests that the GNR-k can be used for nanaoscale heat management and thermoelectric application.

preprint2013arXiv

Superior thermal conductivity and extremely high mechanical strength in polyethylene chains from {\it ab initio} calculation

The upper limit of the thermal conductivity and the mechanical strength are predicted for the polyethylene chain, by performing the {\it ab initio} calculation and applying the quantum mechanical non-equilibrium Green's function approach. Specially, there are two main findings from our calculation: (1). the thermal conductivity can reach a high value of 310 W/K/m in a 100 nm polyethylene chain at room temperature; (2). the Young's modulus in the polyethylene chain is as high as 374.5 GPa, and the polyethylene chain can sustain $32.85% \pm 0.05%$ (ultimate) strain before undergoing structural phase transition into gaseous ethylene.

preprint2012arXiv

A comparative study of two molecular mechanics models based on harmonic potentials

We show that the two molecular mechanics models, the stick-spiral and the beam models, predict considerably different mechanical properties of materials based on energy equivalence. The difference between the two models is independent of the materials since all parameters of the beam model are obtained from the harmonic potentials. We demonstrate this difference for finite width graphene nanoribbons and a single polyethylene chain comparing results of the molecular dynamics (MD) simulations with harmonic potentials and the finite element method with the beam model. We also find that the difference strongly depends on the loading modes, chirality and width of the graphene nanoribbons, and it increases with decreasing width of the nanoribbons under pure bending condition. The maximum difference of the predicted mechanical properties using the two models can exceed 300% in different loading modes. Comparing the two models with the MD results of AIREBO potential, we find that the stick-spiral model overestimates and the beam model underestimates the mechanical properties in narrow armchair graphene nanoribbons under pure bending condition.