This year marks the 100th anniversary of the birth of Professor Zhang Siying. This paper reflects on over three decades of a profound mentor-friend relationship with Mr. Zhang. It expresses admiration for Mr. Zhang’s patriotic feelings, educational demeanor, and scientific research spirit, as well as nostalgia for Mr. Zhang.

Zhang Siying participated in the three-year development of the Hongjian73 control system and made significant contributions. He has successively solved three key problems, among them, identified the key factors that the missile cannot be controled, proposed audacious suggestion for correcting the installation angle of gyroscopes and provided the direction and amount of gyroscope installation angle correction, designed a convenient and intuitive control law based on command waveforms and synthetic vectors to solve the decoupling problem of a type of two channel cross coupling in rotating aircraft. In 1978, he was awarded the "Science and Technology Worker Award for Outstanding Contributions" by the National Science Conference.

After a brief review on the history of semi-tensor product (STP) of matrices, this survey paper introduces general definitions of STP and semi-tensor addition (STA), and the exploring researches on the mathematical essence of STP and STA, including three major branches: Modern Algebra, Geometry, and Analysis. The STP and STA, as cross-dimensional operators, enhance certain developments in classical mathematics, which is basically of fixed dimensions. As a survey paper, it mainly introduces fundamental concepts and basic results with few predictions. We hope to show such a fact that since the STP breaks the dimension barrier of matrix product, it will inevitably cause impact on the classical mathematics, which is of fixed dimensions.

Distributed cooperative localization is a key process in many cooperative tasks for multi-agent systems. Distributed cooperative localization aims to enable each agent to determine their own locations by using the available locations of anchors, the relative measurements to their neighbors, and the distributed network communication. This paper summarizes the advances in the distributed cooperative localization for multi-agent systems. First, according to the types of relative measurements, the distributed cooperative localization methods are classified into three types, i.e., distance-based, bearing-based, and mixed-measurement-based distributed cooperative localization methods. Then, by characterizing the constraint relationship between the measurement and the agents’ locations, the design schemes of the above distributed cooperative localization methods are elaborated in detail, and the advantages and disadvantages of the methods are compared. Moreover, this paper introduces the research status of reliable distributed cooperative localization methods for multi-agent systems with malicious measurements. Finally, the ongoing challenges in distributed cooperative localization are anticipated, and the potential directions for resolution are proposed.

Opinion dynamics has emerged as a research hotspot in many fields such as network science, control theory and sociology, focusing on the analysis and interventions of opinion evolution in complex networks. This paper reviews two research directions developed from the Friedkin-Johnsen model. First, the co-evolution of implicit opinions and explicit opinions under the social pressure is presented, with an emphasis on recent advances in conformity behavior and opinion polarization. Second, the opinion intervention based on opinion maximization problem is introduced and the recent progress is summarized from perspective of intervention strategies such as node selection and timing selection. Finally, future research directions of opinion dynamics under the intersection of multiple fields are discussed.

In order to promote the standardized development of various industries, corresponding standardizing documents need to be formulated in various fields, such as national standard and industry standard. These standardizing documents not only provide a unified operating standard for the industry, but also provide a clear guidance basis for relevant parties. The Central Committee of the CPC and the State Council clearly pointed out in the "the Outlines for the Development of National Standardization" that promoting the digitalization process of standard is an important measure to realize the modernization of the industry. Therefore, it is particularly important to carry out research on the automatic generation of standardizing documents. With the rapid development of artificial intelligence technology, especially the outstanding performance of large language models in text generation tasks, it is possible to use these advanced technologies to realize the automatic generation of standardizing documents. Based on this background, this paper proposes a two-stage scheme for generating standardizing documents. The scheme first generates the outline of the standardizing document through the large model, and then expands to generate the complete document content on this basis. By combining in-context learning and retrieval augmented generation techniques, this method can not only generate high-quality text, but also significantly improve the accuracy and professionalism of the generated content. In order to verify the feasibility of the scheme, we conducted a series of experiments on our self-built dataset, and the results show that the method can effectively generate documents that meet industry standards, and has good practicability and promotion potential.

Aiming at the problem of missing node attributes in graph data, we proposes a novel attribute missing graph learning framework. The framework maps node attributes to edge attributes by reconstructing the structural joint bipartite network. This enables attribute completion and graph tasks to be performed together under a unified framework that can handle both continuous and discrete missing data. According to the attribute homogeneity and structural homogeneity of the attribute graph, we propose an attribute missing representation learning method, which introduces edge embeddings and attention mechanisms to capture the correlations between attribute-attribute and structure-attribute in structural joint bipartite network to enhance the missing attribute learning. Experimental results on four real-world datasets show that our framework outperforms the baselines in both attribute completion tasks, validating the effectiveness of the method.

Human motion generation aims to generate realistic, high-quality human motion. Aiming to summarize the recent advances in text-driven human motion generation technology, through extensively investigating relevant research and literature, this paper systematically reviews the development process and research status of the text-driven human motion generation task. It comprehensively summarizes the model methods related to the task by classifying the generation models and further analyzes the research progress of key technical issues. It summarizes the commonly used datasets and evaluation methods and deeply discusses the unresolved problems and possible future research directions in this field.

Complex criminal cases today present systematic characteristics of multi-factor coupling and dynamic evolution, and their investigation process faces the challenge of nonlinear information integration. Criminal suspects use anti-detection methods such as changing clothes and shoes, facial obstruction, and posture camouflage, combined with complex environmental interference, which significantly reduces the practical effectiveness of single technical means such as face recognition and video structuring. In order to resolve this problem, this article focuses on the actual needs of suspect identification and tracking, breaks through the recognition bottleneck of a single modality, systematically integrates multi-information such as video structuring, face recognition, and gait recognition, and proposes a multi-information fusion video investigation system with gait recognition as the core, which realizes the dual characterization of suspect behavior patterns and identity characteristics, and provides a new technical path for improving identity recognition capabilities and the efficiency of solving complex cases.

In the context of escalating Sino-American strategic competition, a comparative study of Chinese and the USA digital economy policies bears significant strategic value. Traditional methods of policy comparison are limited by cost, can’t solve this problem well. This paper focuses on the contrast between digital economy policies in China and the USA, proposing a resolution framework based on a cross-language model. The framework initially classifies Sino-American digital economic policies by fine-tuning language models and calculating multilingual similarity, thereby achieving automated comparative analysis of the policy environments. Experiments demonstrate that the proposed method can accurately and efficiently identify and extract policy text features, outperforming baseline methods in accuracy across multiple classification dimensions. Finally, by comparing over 16,000 Sino-American digital economic policy texts, this paper reveals key differences in policy tool usage and the focus of digital economic industry development between the two countries, providing a comprehensive and objective portrayal of the disparities in digital economy policy environments. Concurrently, it also brings a fresh perspective to policy comparison research.

This paper investigates the node state estimation of hypergraphs. First, the network model of hypergraphs with pairwise and triplet interactions is built. Second, considering the presence and absence of diffusive coupling, the observer networks are established, and the error dynamical networks are constructed for the two types of hypergraph network models, respectively. Then, using the Lyapunov stability theory, the asymptotic stability of the two types of error dynamical networks is proved and sufficient conditions for state estimation are derived. Finally, the accuracy and effectiveness of the proposed method are verified by numerical simulations. The results indicate the applicability of our method in accurately estimating states within the diffusively coupled and non-diffusively coupled hypergraphs, thereby advancing our capabilities in estimating and controlling higher-order complex networks.

The influence maximization problem in hypergraphs has wide-ranging applications across various fields. Existing methods either inadequately address the redundancy of influence between nodes or only rely on a single metric for initial node ranking, which may fail to accurately capture the true propagation values of nodes. To simultaneously consider both influence redundancy between nodes and the actual propagation values of nodes, this paper proposes a Low Redundant Hypergraph based on the Genetic Algorithm (LR-HGA), which takes into account these two aspects in the selection and crossover operations of genetic algorithm. Experimental results on six real hypergraph networks using the SI propagation model defined on hypergraphs show that the seed set obtained by this algorithm generally has a wider influence spread compared to state-of-the-art benchmark algorithms.

User gender is one of the core aspects of user profiling, and existing methods for accurately identifying user gender mainly rely on user public attributes, with less consideration given to network structure information. This study integrates gender information based on motif theory, subdivides homogeneous motifs into heterogeneous motifs, and proposes a gender recognition method based on heterogeneous motif features, extracting more detailed local information to distinguish users of different genders. Compared to the current popular network embedding methods, the method proposed in this article has improved the Accuracy index by 2.8% to 14.2%, and the AUC index by 2.7% to 15.8%, with stable performance on different proportion training sets. The heteromorphic method can be applied to the identity detection of social users, which helps to conduct in-depth research on the structural characteristics of social networks.

This study aims to address the problem of vital nodes identification in complex networks using graph attention mechanism. This paper integrates both node′s virus transmissibility and structural impact, constructing training labels on the generated network to learn node importance through a graph attention network model. Experimental results demonstrate the excellence of this algorithm in two critical tasks: influence maximization and immune isolation.

For the difficulties of modeling and designing proper controllers for complex network control problems, a model free adaptive control based pinning scheme is proposed to control complex network with unknown and nonlinear coupled relationship in this paper. Firstly, a dynamical linearization model is built based on input/output data of selected pinning node, then a distributed pinning scheme is proposed under minimum variance estimation criterion. This scheme is a data-driven control method because it is designed only with I/O data of pinned nodes instead of network model. The stability analysis for the synchronization error is based on the reduction theorem, contraction mapping method and virtual control. The simulation results demonstrate that the proposed pinning scheme can drive all nodes in network to synchronization states by only control the pinned nodes in network.

To explore the impact of wind power station grid-connection sites on the resilience of power networks, this paper introduces a new analytical framework for assessing the resilience of wind power network. By integrating the network's structural and functional models and applying relevant resilience assessment metrics, we propose a Q-Learning-based grid-connection strategy to identify the optimal grid-connection locations for wind power station. We validate this strategy using the IEEE 118 power grid model, which incorporates wind power grid-connection. Our research shows that the Q-Learning-based grid-connection strategy surpasses traditional heuristic methods and genetic algorithms in reducing operational costs and the risk of overload, highlighting the crucial role of strategic grid-connection in strengthening the network's resilience.

This paper introduces a new type of pseudo-path and incorporate switching signals in the study. We apply graph and matrix theories to explore the controllability of multi-agent systems. Firstly, we obtain the system matrix and its exponential function of the multi-agent system, we then derive the necessary and sufficient conditions to achieve controllability. Secondly, we discuss the impact of different ways of selecting a single leader on the controllability matrix. Finally, we determine the minimum switching period for the system to reach any specified position in the controllable state space under a given switching sequence.

Aiming at the phenomenon that heterogeneous multi-intelligent body systems cannot complete clustering, privacy protection for clustering of heterogeneous multi-intelligent body systems based on the minimum gravitational path is proposed. The differential privacy-preserving Laplace noise is introduced to protect the information privacy of the intelligent body system; the perceptual density algorithm is proposed to improve the adaptability to the initial center intelligent body selection; the gravitational model is constructed, and the Dijkstra algorithm is used to compute the minimum gravitational path, which ensures that all the intelligences can be assigned to the corresponding groups. Experiments show that the algorithm successfully completes the clustering analysis of the heterogeneous multi-intelligent body system, and at the same time guarantees the statistical characteristics of the intelligent body data and realizes the privacy protection.

To explore the driving effect of incentive mechanisms on residents′ enthusiasm for participating in garbage classification, this article is based on empirical data analysis of garbage classification in multiple communities. Using the RFM model, it quantifies residents′ enthusiasm for participating in garbage classification and categorizes user value. Based on this, a multi-agent simulation model of user value transformation is established around subjective norms, classification knowledge, and classification attitudes. The results show that residents′ enthusiasm for garbage disposal is influenced by subjective norms; there is a positive correlation between classification knowledge, classification attitudes, and residents′ enthusiasm for garbage disposal.