Journal of Zhengzhou University(Natural Science Edition)

 

Current Issue

  • A Survey on the Construction and Application of Knowledge Graphs Driven by Large Language Models

    ZHANG Kunli;WANG Ying;FU Wenhui;ZHU Yongqi;ZHANG Yanli;ZAN Hongying;School of Computer and Artificial Intelligence, Zhengzhou University;School of Cyber Science and Engineering, Zhengzhou University;

    With the support of large language models, knowledge graphs, characterized by their structured and semantically rich features, enhanced data association and interpretability. This provided new research directions and application potential for complex knowledge reasoning and intelligent decision-making. Therefore, from the perspective of knowledge graphs, the latest progress in the construction and application of knowledge graphs driven by large language models was summarized. Firstly, the new methods for building knowledge graphs were explored from the perspectives of knowledge modeling, information extraction, knowledge integration, and knowledge graph completion. Secondly, the applications of knowledge graphs in augmenting large language models, improving retrieval capabilities, and achieving mutual enhancement with large language models were investigated. Finally, the future directions of combining large language models with knowledge graphs were discussed.

    2026 02 v.58 [Abstract][OnlineView][Download 1046K]

  • A Prototype Optimization and Refinement Segmentation Network for Few-shot Medical Image Segmentation

    WEI Mingjun;HE Haipeng;CHEN Weibin;LIU Yazhi;LI Hui;College of Artificial Intelligence, North China University of Science and Technology;Hebei Provincial Key Laboratory of Industrial Intelligent Perception;Imaging Center, North China University of Science and Technology, Affiliated Hospital;

    In order to solve the problem of few-shot medical image segmentation with the distribution shift and local edge details between the support set and the query set, a prototype optimization and refinement segmentation network(PORSNet) for few-shot medical image segmentation was proposed. The network contained a prototype loop iteration module, which suppressed the distribution shift between the initial prototype and the query set and enhanced the expressiveness of the prototype by iteratively performing steps such as initial prototype correction, prototype global perception, and prototype distillation. In addition, a prototype refinement segmentation module was included to further process edge detail information through mask-guided aggregation and feature normalization refinement. Extensive experiments on two widely used few-shot medical public datasets, ABD-MRI and ABD-CT, showed that the proposed PORSNet could use a small number of samples to ensure the segmentation effect with good generalization ability.

    2026 02 v.58 [Abstract][OnlineView][Download 1319K]

  • Bus Passenger Flow Prediction Based on Multiple Information Attention and Adversarial Graph Convolution

    YAN Jianqiang;ZHAO Renqi;GAO Yuan;QU Boting;School of Information Science and Technology, Northwest University;School of Economics and Management, Northwest University;

    Aiming at the difficulty of utilizing spatiotemporal dependence relationship in bus passenger flow prediction effectively, a prediction model of passenger flow based on multiple information attention and dynamic adaptive adversarial graph convolutional network was proposed. Firstly, the time feature encoder was used to capture the similarity between passenger flows at different time periods, and point of interest(POI) information of bus stations was incorporated to enhance node feature extraction. Secondly, the dynamic modeling of spatiotemporal dependence was adopted to complete the modeling of non-Euclidean relationships, and the SimAM attention module was utilized to capture the overall differences in passenger flow data at different stations. The experimental results on real bus passenger flow data showed that compared with the best baseline model, the proposed model reduced the average MAE and RMSE of the next 12 time steps by 0.34 and 0.33, respectively, demonstrating its effectiveness and superiority in predicting bus passenger flow.

    2026 02 v.58 [Abstract][OnlineView][Download 1410K]

  • Domain Adaptation Based on Prompt Learning and Classification Certainty Maximization

    DING Meirong;ZHUO Jinxin;LIU Qinglong;LANG Jicong;School of Artificial Intelligence, South China Normal University;

    Domain adaptation faced the issue of complex and variable real-world scenarios, and existing methods mostly focused on optimizing classification consistency while neglecting classification certainty. To address these issues, a network model combining constrastive language-image pre-training(CLIP) with classification certainty maximization was proposed. CLIP, as a multimodal pre-trained model, was pre-trained on a large scale of image-text pairs and possessed strong cross-domain generalization capabilities. By leveraging prompt learning and contrastive learning, the knowledge of the CLIP model was acquired, enabling the model to adapt more complex real-world scenarios. Through the method of classification certainty maximization, a dual-classifier was employed to assess classification consistency and reduce confusion during the model′s inference process. Experiments were conducted on three domain adaptation benchmark datasets: Office-31, Office-Home, and MiniDomainNet. The experimental results indicated that compared with existing advanced methods, the proposed model showed improvements in image classification accuracy across all three datasets.

    2026 02 v.58 [Abstract][OnlineView][Download 1043K]

  • A Reinforcement Learning Based Approach to Multi-objective Microservice Deployment

    ZHANG Menyao;ZHANG Yingxi;ZHENG Wenqi;FENG Guangsheng;School of Computer Science and Technology, Harbin Engineering University;

    In edge computing, microservice architecture could improve data processing efficiency and application response speed, which was suitable for various application scenarios with fast response and frequent interactions. However, existing studies neglected the impact of different interaction frequencies between microservices on the communication overhead. To address this problem, a multi-objective microservice optimal deployment method based on reinforcement learning to improve the performance of microservices in edge environments was proposed. A dual optimization objective model that considers reducing the communication overhead of microservice interactions and balancing the resources of edge nodes was established. Then a deep Q-learning algorithm based on an improved reward mechanism was designed. In order to adapt to the characteristics of shared resources in the process of microservice deployment, a shared reward mechanism was introduced so that the algorithm had better convergence. The experimental results showed that the proposed algorithm could balance the microservice interaction perception and node resource utilization better, and had shorter response time compared with the existing DIM method and Kubernetes default deployment method.

    2026 02 v.58 [Abstract][OnlineView][Download 1534K]

  • Construction of a Scenarioized Knowledge Graph of International Chinese Teaching Resources

    YANG Hao;XIN Jing;ZHU Shanyi;RAO Gaoqi;XUN Endong;Beijing Advanced Innovation Center for Language Resources,Beijing Language and Culture University;School of Chinese Language and Literature,Beijing Foreign Studies University;The School of International Education,Shandong Normal University;Research Institution of International Chinese Language Education,Beijing Language and Culture University;

    In recent years, numerous knowledge bases were developed by the academic community to support international Chinese language teaching. However, nost of them targeted at specific resources, such as collocation databases or example sentence collections, resulted in isolation. In the era of ubiquitous intelligence, international Chinese education had to face with the challenge of a digital and intelligent paradigm, with higher demands on language teaching resources. The construction of a fine-grained knowledge graph with various resource entities was considered essential. In educational contexts, particular emphasis was placed on "teaching according to the student′s aptitude". The construction of knowledge graphs for teaching should take into account the provenance and application of knowledge, i.e. scenario-based relevance. The BCC structural retrieval tool was employed to link various resource entities, with careful consideration given to the origin of knowledge and its applicable contexts. As a result, a contextualized knowledge graph for international Chinese language teaching was constructed, and preliminary experiments practices were conducted to explore its application in intelligent international Chinese language education.

    2026 02 v.58 [Abstract][OnlineView][Download 1099K]

  • Formal Analysis of ID-AOFE Protocol Based on Time-event Logic

    XIAO Meihua;QIAO Shanshan;YANG Ke;School of Information and Software Engineering, East China Jiaotong of University;

    Fair exchange protocols aimed at providing a secure and fair mechanism for digital information exchange. Analyzing the fairness of such protocols was an important research content in the field of information security. Time-event logic had a mechanism to describe the knowledge and state changes of protocol subjects over time and was considered an effective method for analyzing the security attributes of protocols. Based on time-event logic, in view of the characteristics of mutual distrust and deceptive behaviors in fair exchange protocols, the fairness of the protocol was analyzed by determining whether there was a strategy that could enable dishonest subjects to gain additional advantages when the protocol ended its operation. An example was analized by using an identity based-ambiguous optimistic fair exchange(ID-AOFE) protocol. A standardized message interaction process was defined. A fine-grained analysis of the time in the message interaction process of the protocol was conducted. Two fairness loopholes in the protocol were discovered. The entire process of the attack occurrence was given in combination with the graphical description method, illustrating the validity of the theory.

    2026 02 v.58 [Abstract][OnlineView][Download 905K]

  • Weapon Target Assignment Based on Adaptive Tabu Search Multi-objective Whale Optimization Algorithm

    ZAI Guangjun;XU Wangwang;ZHONG Lihong;TIAN Zhao;SHE Wei;School of Cyber Science and Engineering, Zhengzhou University;Songshan Laboratory;Zhengzhou Key Laboratory of Blockchain and Data Intelligence;

    Aiming at the problems of empirical parameter setting, insufficient population diversity, and weak searchability of the multi-objective whale optimization algorithm for solving the weapon target allocation problem, an adaptive tabu search multi-objective whale optimization algorithm was proposed. Firstly, adaptive grid partitioning and external archiving adjustment strategies were adopted to allow the grid and archive sizes to be automatically adjusted based on population distribution and diversity changes. Secondly, a dynamic roulette wheel selection method was designed to control the generation of global optimal individuals, thereby enhancing diversity and uniformity of population distribution. Additionally, the tabu list and neighborhood search strategies from the tabu search algorithm were introduced to expand the population′s exploration capability in new areas. Simulation experiment results demonstrated that the proposed algorithm exhibited superior population distribution and solution set diversity along with faster-solving efficiency, effectively improved quality of the solution set, and could better solve the multi-objective weapon allocation optimization problems.

    2026 02 v.58 [Abstract][OnlineView][Download 1172K]

  • A Location Selection Method of Charging Station for Electric Vehicles Based on Improved NSGA-Ⅱ Algorithm

    YU Ning;FENG Xin;TANG Aihua;SHU Zirong;School of Mechanical Engineering, Chongqing University of Technology;School of Vehicle Engineering, Chongqing University of Technology;

    The strategic placement of electric vehicle charging stations is of utmost importance in the advancement of electric vehicles. The unreasonable arrangement of charging stations could result in increased expenses of operators and diminished satisfaction of users. To tackle these challenges, a synthetical model was constructed to quantify total operating cost and user satisfaction of charging stations. In this model, the total operator costs were segmented into land cost, construction cost, operating cost and government subsidy. The user satisfaction was quantified by charging distance and waiting time. The improved non-dominated sorting genetic algorithm Ⅱ(NSGA-Ⅱ) was designed to address the multi-objective optimization model. The solution solved the problem of minimizing the operating cost and maximizing the user satisfaction with the constraints of remaining power, charging distance, and the quantity of charging piles. In the end, Jiangbei was taken as a case study to simulate the location selection of electric vehicle charging stations, confirming the efficacy of the proposed approach.

    2026 02 v.58 [Abstract][OnlineView][Download 917K]

  • Kinetic Reaction Mechanism of Methyl-triggered Phenylacetylene to Form Bicyclic Aromatic Indene

    FENG Shihao;WANG Changyang;ZHANG Lijuan;ZHAO Long;JIANG Huiling;BIAN Huiting;School of Mechanics and Safety Engineering, Zhengzhou University;National Synchrotron Radiation Laboratory, University of Science and Technology of China;Technical Support Base for Prevention and Control of Major Accident in Metal Smelting, University of Science and Technology Beijing;

    In order to reveal the formation mechanism of the bicyclic aromatic hydrocarbon indene, the reaction potential energy surface of methyl radical and phenylacetylene was studied, and the kinetic data were calculated. Firstly, the B3LYP/6-311+G(d, p)method was used to optimize the geometry and analyze the vibration frequency of all species in the reaction system of phenylacetylene and methyl radical. Then, the CCSD(T)/aug-cc-pVDZ method was used to obtain high-precision energy of each species. Finally, reaction rate constants and branching ratios of important paths were ultimately computed by the time-dependent RRKM/ME theory. Theoretical results indicated that the addition of methyl radical onto the acetylenyl group in phenylacetylene would trigger the subsequent isomerization and dissociation. And on variable pressure(0.001～0.1 MPa)and temperature(300～3 000 K)conditions, 1-phenyl-1-propyne(C_9H_8)+H were invariably the main product for current system. Moreover, the low pressures conduced to yield indene, and its formation were competing fiercely for the bimolecular products, i.e. C_6H_5+C_3H_4, C_6H_5+C_3H_4-2 and Int41-P4+C_2H_2, while high pressures were in favor of these bimolecular products generated.

    2026 02 v.58 [Abstract][OnlineView][Download 1882K]

  • Study on the Dynamic Model of Three-layer Composite Beams for Steel Spring Floating Plate Track System

    XU Ping;ZHANG Yunpeng;WANG Weidong;YANG Yanfeng;LIU Xiangming;School of Water Conservancy and Transportation, Zhengzhou University;China Academy of Railway Sciences Group Co.,Ltd.;China Academy of Railway Sciences (Shenzhen) Research and Design Institute Co., Ltd.;

    With the rapid development of urban rail transit, the negative impact of vibration and noise on the environment was increasingly significant, and the research on steel spring floating plate track vibration reduction technology was of great significance. The differences of structural parameters of steel rails, fasteners, floating slabs, steel spring isolators, concrete bases, and foundation were fully considered, the steel spring floating slab track system was simplified as a three-layer Euler Bernoulli beam, the dynamic equations with moving loads were constructed, and the analytical solution of the three-layer composite beam model was obtained with double Fourier transform. The maximum force transmitted to the base was adopted to characterize the vibration reduction performance of the track system; the vertical displacement of the steel rail was adopted to characterize the safety and comfort performance of train operation; and the influence of various factors on the vibration reduction performance of the track system was analyzed. The reduce of the support stiffness of the steel spring could improve the vibration reduction performance of the track system, but it would increase the peak vertical displacement of the steel rail and reduce the speed at which the displacement peak occurs. The increase of the support damping of steel springs within certain range could improve the vibration reduction performance of track systems. The increase of the mass of floating slabs could improve the overall strength of the track system, but it would have certain adverse effects on the vibration reduction performance of the track system. When factors such as vibration reduction performance, comfort performance, structural safety and durability of the track system were taken into account, recommendations for parameter values of steel spring stiffness, damping, and floating plate mass were provided. They were(4～8)×10~6 N/m,(6～9)×10~4 N·s/m, and(1 600～2 500) kg/m respoctively.

    2026 02 v.58 [Abstract][OnlineView][Download 1689K]

  • Research on the Publication Trend of Natural Science Edition of Journals of Double First-class Universities from 2015 to 2024

    WANG Haike;Archives and School History Museum of Zhengzhou University;

    University journal is a special type of adademic journal with general dilemma of lacking high-quality manuscript sources and up-to-date pulishing concepts in today′s development. The academic influence of the natural science edition of 42 double first-class universities journals increased significantly between 2015—2024, and their publication pattern could cast some extra light on the development of university journal. Twenty basic descipline journals of them were selected as the target of this study. Data of four key publication indicators of these university journal were collected and analyzed. From the findings of this study, the following conclusions were drawn. For each journal, it was compulsory to optimize the number of articles with its status. The number of articles published should match its academic level. The normal citation of reference should be treated with respect, abnormal citeations should be rejected. The objective development law of average count of citation should be followed. All high quality articles could be accepted and published. The concept of funded article only should be abandoned. The composite impact factor indicators should be treated objectively and used scientifically. It should be compared with historical data or data of similar journal. University journal should follow these rules and take into account other factors, such as columns, disciplines, publication cycles, and international communication to find a suitable path for its own development.

    2026 02 v.58 [Abstract][OnlineView][Download 1105K]