中国科学院数学与系统科学研究院期刊网

14 April 2025, Volume 45 Issue 4
    

  • Select all
    |
  • TIAN Peiyu, WANG Xihui, FAN Yu, ZHU Anqi
    Journal of Systems Science and Mathematical Sciences. 2025, 45(4): 994-1012. https://doi.org/10.12341/jssms240027
    Abstract & Html ( ) Download PDF ( )   Knowledge map   Save
    In recent years, there have been more frequent disasters occurred in China, which pose significant threats to the lives and property of the people. To cope with the increasing complexity and severity of disasters, decision-makers need to store and dispatch emergency supplies rationally based on the real situation. Current studies on regional dispatch considering multiple warehouses and demand points are insufficient, and the problems such as ‘who/how/how much to dispatch’ have not been well-answered. To solve these problems, this paper proposes three regional dispatching strategies (including strict administrative hierarchy supply dispatch, cross-administrative hierarchy supply dispatch and free and nearest supply dispatch strategies) based on a comprehensive summary of relevant case studies, then builds a multi-agent simulation model based on deprivation cost. A simulation experiment is conducted in Mengcheng County, Bozhou City, Anhui Province, and the result shows that when the regional demand is large in a short time, the free proximity strategy can minimize the total social logistics cost. On the contrary, when the regional demand is small, the differences of the total social cost among three strategies are small. In conclusion, our research suggests that, when facing severe disasters and catastrophes, governments should cooperate and coordinate on the dispatching of relief supplies. However, when facing normal disasters without the risk of life, the demand can be satisfied with the strict administrative strategy.
  • YAO Yitao, JIA Bin, ZHAO Tingting
    Journal of Systems Science and Mathematical Sciences. 2025, 45(4): 1013-1030. https://doi.org/10.12341/jssms240089
    Abstract & Html ( ) Download PDF ( )   Knowledge map   Save
    Identifying key segments within road networks is crucial for selecting repair objectives and optimizing repair sequences during the post-disaster recovery phase. Traditional methods for identifying key segments have not fully explored the interactions between multiple segments, particularly the significance of studying road network vulnerability under simultaneous disruptions of multiple links. To tackle this issue, this study introduces a machine learning model called transportation graph attention networks for criticality analysis (TGAT) to identify key road segments when facing multiple disruptions. This model is trained on data samples that include scenarios of multiple segment failures, utilizing the graph attention network to evaluate the influence weights between segments and calculating the criticality of each segment based on these weights. The model, trained using mean squared error as the loss function, is capable of identifying segments that play a crucial role in the performance of the road network. Taking the Kunshan City road network as an example, this paper compares the effectiveness of the TGAT method with three other methods:Degree centrality, weighted betweenness centrality, and eigenvector centrality, in optimizing repair sequences during the post-recovery phase. Experimental results indicate that the TGAT method is more effective in identifying key segments within the road network compared to the other three methods, and the repair sequence optimized using TGAT further enhances the repair performance of the road network.
  • FENG Jiawei, DAI Bitao, BU Tianci, ZHANG Xiaoyu, OU Chaomin, LÜ Xin
    Journal of Systems Science and Mathematical Sciences. 2025, 45(4): 1031-1043. https://doi.org/10.12341/jssms240058
    Abstract & Html ( ) Download PDF ( )   Knowledge map   Save
    In the numerous terrorist attacks that have occurred worldwide, various terrorist organizations have shown a trend of collaborative cooperation, posing significant challenges to international counter-terrorism efforts. Based on the global terrorism database (GTD), this study constructs a terrorist organization cooperation evolution network from 121,074 terrorist attacks that occurred globally from 2001 to 2018 and conducts a time-series topological structure analysis. Based on the characteristics of terrorist organization cooperation, the network is divided into time slices of three years each to model the flow patterns of terrorist communities at multiple scales. The analysis shows that the robustness of the terrorist organization cooperation network has been continually strengthening over time, which is necessary to develop corresponding strategies to disrupt it. Focusing on the largest connected sub-network within the terrorist cooperation network, whose influence is continuously expanding, this study proposes a community structure-based neighborhood centrality index (CSNC) to measure the importance of nodes in the largest connected component. Experimental results demonstrate that the network disruption strategy based on CSNC, in the process of disintegrating the terrorist cooperation network from 2001 to 2018, achieved a 16.45% maximum reduction in the R value compared to benchmark strategies, proving that the CSNC-based disruption strategy can more effectively dismantle terrorist cooperation networks.
  • ZHU Li, YANG Yaoxing, CHU Deshui, HU Chenke
    Journal of Systems Science and Mathematical Sciences. 2025, 45(4): 1044-1063. https://doi.org/10.12341/jssms240090
    Abstract & Html ( ) Download PDF ( )   Knowledge map   Save
    Exploring the heterogeneity characteristics of nodes and edges in emergency logistics networks has a great impact on the location-allocation optimization problem throughout the entire emergency system. First, for the disaster preparedness stage, based on the complex network theory, this paper selects the node vulnerability indicators to characterize the heterogeneity of various emergency reserve facilities. The vulnerability indicators of transportation connected edges among emergency reserve facilities, and between emergency reserve facilities and disaster-affected demand areas, are also incorporated into the emergency location-allocation optimization decision-making. And a multi-objective location-allocation model is formulated, for comprehensively balancing vulnerability, efficiency and cost-effectiveness in the emergency logistics network. Then, taking the emergency reserve network covering 13 prefecture-level cities in Jiangsu Province as a case example, an NSGA-II algorithm is designed to solve the constructed model and perform a numerical simulation. Not only the sensitivity analysis for the parameters is analyzed, but also the comparative analysis is discussed. The simulation results show that the location-allocation optimization problem for heterogeneous emergency reserve facilities from the perspective of vulnerability, not only performs well in cost-effectiveness and allocation efficiency, but also significantly improves the robustness and resilience of the entire emergency logistics network. This work provides some useful managerial insights for emergency decision-makers to make an effective disaster preparedness plan.
  • ZHAI Weinan, DING Ying, YU Jianjun, ZHANG Lingling
    Journal of Systems Science and Mathematical Sciences. 2025, 45(4): 1064-1081. https://doi.org/10.12341/jssms240130
    Abstract & Html ( ) Download PDF ( )   Knowledge map   Save
    Currently data elements as an important strategic resource for enterprises, data security is the basis for the survival and development of enterprises, the theft, leakage, tampering, destruction, abuse of data and other issues, will bring serious threats and damage to enterprises. This paper focuses on solving the problem of security risk assessment based on enterprise data assets, constructing risk evaluation indexes from four aspects of assets, vulnerability, threat, and protection, and proposing to realize the independence analysis between different indexes based on principal component analysis. Meanwhile, considering the correlation characteristics between data assets, we design a multi-asset correlation analysis enterprise data risk assessment model based on the comprehensive gray correlation, which effectively solves the problem of repeated risk assessment of correlated assets, improves the accuracy of risk assessment of enterprise data assets, and provides decision-making suggestions for the security protection of enterprise data.
  • GAO Dayou, YANG Kai, YANG Lixing, HAO Yuchi, WANG Entai
    Journal of Systems Science and Mathematical Sciences. 2025, 45(4): 1082-1101. https://doi.org/10.12341/jssms240382
    Abstract & Html ( ) Download PDF ( )   Knowledge map   Save
    In view of the huge amount of excavation earthwork, strong uncertainty and dynamic characteristics of earthwork allocation in major projects, the comprehensive utilization mode of earthwork and the dynamic siting strategy of the consumption sites are proposed. To maximize the expected total profit of the project, a two-stage stochastic programming model which considers the excavation sequence, machinery efficiency and the uncertainty of excavation amount is established. According to the structural characteristics of the proposed model, an improved Benders decomposition algorithm combined with sample average approximation method is designed, and two acceleration strategies are introduced to speed up the convergence of the developed algorithm. Finally, different scale numerical experiments are conducted for testing. The experimental results show that the proposed comprehensive utilization method and the dynamic siting strategy of consumption sites can improve the utilization value and reduce the cost of earthwork. The two-stage stochastic programming model can effectively characterize the uncertain excavation amount, and the designed accurate algorithm can quickly solve the problem. The research results can provide decision basis and algorithm support for making the dynamic location plans of the consumption sites and earthwork allocation schemes for the major projects.
  • YAN Shuai, LI Yimin
    Journal of Systems Science and Mathematical Sciences. 2025, 45(4): 1102-1115. https://doi.org/10.12341/jssms23661
    Abstract & Html ( ) Download PDF ( )   Knowledge map   Save
    In this paper, we analyze the pricing, purchasing decision and inventory information disclosure strategy of an online retailer in the presence of the conspicuous consumption consumers. In particular, we investigate two scenarios in which the online retailer either discloses or conceals inventory information. Our major findings are summarized as follows. First, we determine and quantify that the proportion of conspicuous consumers has significant profit implications for the online retailer, and observe variations in this implication under different inventory information disclosure strategies. Second, the disclosure strategy of inventory information is primarily determined by the proportion of conspicuous consumers and their sensitivity to product sales. Finally, the hassle cost of regular consumers has a certain impact on the retailer's inventory information disclosure strategy. The insights we generate can provide guidance to online retailers in making informed decisions to achieve specific objectives.
  • LIU Changshi, LI Junyu, ZHAO Shen, ZHOU Xiancheng, FAN Lijun
    Journal of Systems Science and Mathematical Sciences. 2025, 45(4): 1116-1139. https://doi.org/10.12341/jssms240031
    Abstract & Html ( ) Download PDF ( )   Knowledge map   Save
    The advent of renewable energy charging stations and vehicle-to-grid (V2G) technology heralds unprecedented possibilities for the logistics sector. Within the distribution framework, electric vehicles can funnel a portion of the energy acquired from these renewable energy charging stations back into grid-connected thermal power stations via V2G technology, thereby generating significant V2G revenue. By accounting for factors such as the energy heterogeneity across charging stations, customer demand, electric vehicle energy consumption, carbon emissions, and potential V2G profits, a comprehensive mixed-integer programming (MIP) model for electric vehicle distribution, charging, and electricity transmission has been developed. This model aims to minimize the net discrepancy between the total distribution costs and V2G revenues. To address the problem's unique challenges, a hybrid ant colony algorithm (HACA) has been engineered. Numerical experiments employing multiple types of instances substantiate the efficacy of the proposed methodology. The findings reveal that the proposed approaches not only substantially curtail overall distribution expenses while augmenting V2G profits but also achieve "zero-emission" distribution for electric vehicles. Moreover, the proposed approaches offer cost-effective avenues for integrating renewable energy into the grid, fostering a synergistic, mutually beneficial relationship among logistics firms, utility companies and end-users.
  • XU Linming, YI Sicheng
    Journal of Systems Science and Mathematical Sciences. 2025, 45(4): 1140-1155. https://doi.org/10.12341/jssms240268
    Abstract & Html ( ) Download PDF ( )   Knowledge map   Save
    In order to understand the overall advantages and disadvantages of each alternative in a certain period of time, it is necessary to conduct dynamic evaluation and analysis. Aiming at the large group utility value of the VIKOR, which only considers the distance between alternatives and the ideal solution at the single moment, this paper introduces the distance between the connection vector to obtain a more reasonable group utility value, and absorbs the idea of grey correlation analysis to evaluate the shape similarity between alternatives and the ideal solution, finally, a dynamic grey correlation VIKOR method based on connection vector distance is proposed by incorporating time factors. This method fully considers maximizing group utility and minimizing individual regret, as well as the geometric similarity between the data curve of alternatives and the ideal solution. At the same time, it also takes into account the degree of difference and growth of indicator values, and the corresponding parameter values can be adjusted according to the preferences of decision-makers. Finally, the method is applied to the evaluation of headquarters economic development level in 13 eastern cities of China from 2015 to 2019 to verify the scientificalness and effectiveness of this method, and its effectiveness is further verified through comparison with other methods.
  • WAN Jianxiang, NIE Changteng
    Journal of Systems Science and Mathematical Sciences. 2025, 45(4): 1156-1176. https://doi.org/10.12341/jssms23725
    Abstract & Html ( ) Download PDF ( )   Knowledge map   Save
    Based on the two phases of the China Household Finance Survey(CHFS), the difference-in-differences and difference-in-difference-in-differences methods are used to study the impact of digital access on rural household consumption from the perspective of bridging "access divide" and "use divide". The research results find that: linebreak 1) The digital access has a significant positive impact on the rural household consumption, and it is mainly reflected in the promotion of the growth of rural households' subsistence and enjoyment consumption, but it has not improved developmental consumption and promoted the upgrading of household consumption. 2) Both the "access divide" and the "use divide" have promoted the growth of rural household consumption, and bridging "use divide" is greater in playing a role. 3) Heterogeneity analysis shows that digital access has a greater positive impact on the consumption of households in the educational group, households whose heads of household have reached secondary education level, high income households, and rural households in the eastern region. 4) The mechanism is that the digital access reduces the cost of information search, activates entrepreneurial activities and increases household income levels, thereby promoting the growth of rural household consumption. Finally, it is recommended to expand the coverage of network infrastructure in rural areas, focus on improving the education level of rural residents and carry out digital skills training to help farmers share the dividends of digital economy development.
  • ZHU Furong, YUE Dequan
    Journal of Systems Science and Mathematical Sciences. 2025, 45(4): 1177-1192. https://doi.org/10.12341/jssms23605
    Abstract & Html ( ) Download PDF ( )   Knowledge map   Save
    In this paper, we consider a M/M/c queueing-inventory system under lost sales. It is assumed that all servers take multiple vacations synchronously when the inventory is zero, and the vacation time is assumed to follow a general distribution. The system adopts (r, Q) policy, and the lead time is assumed to be exponentially distributed. Firstly, we consider a case that the service time is negligible. For this case, the steady-state probability distribution of the inventory level in the system is obtained by using the supplementary variable method. On this basis, the M/M/c queueing-inventory system with positive service time is considered, and the product form solution of steady-state joint probability distribution of queue length, the on-hand inventory level and servers' status are obtained. Furthermore, some performance measures of the system such as average inventory level, average order rate and average customer loss rate are obtained. Finally, the effects of system parameters on some important performance measures are analyzed by numerical examples. In addition, the average cost model per unit time of the system is established, then the optimal replenishment strategy, the optimal number of servers and the optimal average cost of the system are calculated by using the genetic algorithm.
  • TIAN Ruiling, FU Yueyang
    Journal of Systems Science and Mathematical Sciences. 2025, 45(4): 1193-1212. https://doi.org/10.12341/jssms23925
    Abstract & Html ( ) Download PDF ( )   Knowledge map   Save
    This paper investigates the equilibrium strategies of customer in single-server Markovian queues with multiple vacations and catastrophes. Whenever a catastrophe occurs in the system, all customers in the queue are forced to leave, which causes the server to fail. And customers are not allowed to enter the system during the repair period. The steady-state probabilities of the system are obtained under different situations:Fully observable, almost observable, almost unobservable, and fully unobservable. Based on the reward-cost structure, we obtain the customer equilibrium strategies and social benefits in four situations. In addition, we provide some numerical experiments to illustrate the effects of information level and system parameters on the equilibrium behavior of customers.
  • XIAO Huimin, HU Yada
    Journal of Systems Science and Mathematical Sciences. 2025, 45(4): 1213-1228. https://doi.org/10.12341/jssms240056
    Abstract & Html ( ) Download PDF ( )   Knowledge map   Save
    Fuzzy logic expresses preference in uncertain language and increases the flexibility and practicability of decision-making process by dealing with uncertainty. It has become a research hotspot in decision-making field in recent years. In the past decision-making methods, the relative value of each scheme is given by the complete ranking, or the aggregative ranking is given according to the classification information, while the complex fuzzy decision-making problem should be made from a comprehensive multi-dimensional perspective. In this paper, the rank-sum ratio method (RSR) is introduced into the field of fuzzy decision-making, and a comprehensive ranking method based on Pythagoras fuzzy numbers is proposed. Based on the prospect theory, the prospect synthesis matrix for scheme evaluation is established, and then the prospect matrix is ranked, and the complete ranking of schemes is sorted by the rank sum ratio. Finally, the data is mapped to the normal distribution curve according to various cases of rank, and the aggregative ranking is carried out according to the correlation partitioning method of normal distribution. The effectiveness and superiority of the algorithm are verified by comparative analysis. The purpose of this study is to provide a new idea reference for the application of rank sum ratio method in the field of fuzzy decision-making.
  • DENG Bing, ZHANG Xingong
    Journal of Systems Science and Mathematical Sciences. 2025, 45(4): 1229-1241. https://doi.org/10.12341/jssms240017
    Abstract & Html ( ) Download PDF ( )   Knowledge map   Save
    In this paper, the rejection problems of the two-agent single machine under generalized parameters are studied. Generalized parameters refer to the parameters that are given in advance according to the position of the job in sequence. Under the three-parameter types of agent jobs with generalized weight(GW), generalized rejection cost (GRC) and generalized processing time(GPT), the jobs of the first agent can be rejected, and its objective function is to minimize the sum of the total weighted completion time and the total rejection cost. The maximum cost function of the second agent does not exceed the fixed value. We find a schedule to minimize the objective functions of the first agent. We provide NP-hardness proofs for the generalized weight and generalized rejection cost problems, and pseudo-polynomial time optimal algorithms. We also provide a polynomial time optimal algorithm for the generalized processing time problem. Finally, we give the case experiments.
  • LIU Lifeng, YAN Xingyu, ZHANG Xinyu
    Journal of Systems Science and Mathematical Sciences. 2025, 45(4): 1242-1254. https://doi.org/10.12341/jssms240096
    Abstract & Html ( ) Download PDF ( )   Knowledge map   Save
    Currently, one of the main challenges in practical modeling lies in the fact that training and testing data come from different distributions. Stable learning addresses this issue by decorrelating all covariates through sample reweighting, thereby achieving stable predictive performance. While machine learning methods such as stable learning show good results in experiments, there are still theoretical gaps, such as the lack of metrics for model stability under testing data and explanations for why stable learning maintains stable predictions across multiple environments. This paper proposes a new metric of stability, compares stable learning methods with ordinary least squares and explores the reasons why stable learning maintains stability across multiple environments. Finally, the paper validates the theory through simulated experiments. This research contributes to refining the theory of stability in stable learning, enhancing the understanding of stability in stable learning, and guiding the selection of practical modeling methods.
  • CHEN Qianru, HE Jianfeng
    Journal of Systems Science and Mathematical Sciences. 2025, 45(4): 1255-1278. https://doi.org/10.12341/jssms23642
    Abstract & Html ( ) Download PDF ( )   Knowledge map   Save
    In the context of big data, it is easier to obtain more available auxiliary information. However, due to the high dimensional characteristics of auxiliary information, the computational complexity increases and the estimation effect of superpopulation model is reduced, thus making the traditional model-assisted estimation method no longer applicable. Therefore, in the framework of model-assisted sampling estimation, this paper introduces the Bayesian model averaging method to estimate the superpopulation model, and obtains the combined model auxiliary estimator under the complete auxiliary information. Firstly, the traditional model-assisted sampling estimation method and Bayesian model averaging method are systematically sorted out. Then the Bayesian model averaging assisted sampling estimation method is given and the asymptotic unbiasedness and consistency of the estimator are proved. The numerical simulation and empirical analysis show that the Bayesian model averaging assisted estimator is better than the generalized regression estimator using a single model in the presence of high-dimensional auxiliary information. Finally, based on the research summary, the prospect of improving the Bayesian model averaging assisted sampling estimation method is proposed.
  • XIAO Han, WANG Chunjie, XU Ping, YU Dan, WANG Di
    Journal of Systems Science and Mathematical Sciences. 2025, 45(4): 1279-1294. https://doi.org/10.12341/jssms23337
    Abstract & Html ( ) Download PDF ( )   Knowledge map   Save
    In this paper, we propose a varying-coefficient functional linear quantile regression model with right censored data. We use functional principal component analysis and B-spline methods to reduce the dimensionality of functional coefficients and varying-coefficient functions, respectively. The estimation of unknown functions is obtained by minimizing a weighted quantile loss function, and the large-sample properties of the estimator are obtained under certain regularization conditions. The feasibility of the model is demonstrated by some simulation studies, and the practicability of the model is illustrated by laryngoscope image data.
  • LIU Xiaohui, DONG Yuxin, CHENG Hongbo, QIU Jie
    Journal of Systems Science and Mathematical Sciences. 2025, 45(4): 1295-1307. https://doi.org/10.12341/jssms23880
    Abstract & Html ( ) Download PDF ( )   Knowledge map   Save
    Transformers are important equipment in power systems, and monitoring their operating status and accurately identifying possible faults is of great importance. However, there are many indicators in the transformer oil chromatographic data, and the evaluation method that only relies on a single indicator does not consider the correlation and non-normality of the indicators, which will affect the correctness of the evaluation results. In view of this, this paper combines the non-parametric tool of projected outlier function to construct a new class of transformer status hierarchical early warning method for multi-index data. The new method does not need to assume the distribution of oil chromatography data in advance. The different identification windows it creates can calculate the transformer state projection outlier(AO) value at different times in each window. Based on the AO value, the transformer state type can be divided, thereby realizing the transformation of the transformer. Status hierarchical warning. The results show that the new method effectively solves the impact of correlation and non-normality on the monitoring performance of the model to a certain extent, and the model has the characteristics of achieving hierarchical early warning, easy visualization and high recognition sensitivity.
  • MENG Jie, QIAO Tingting, GENG Chenbo, YANG Guijun
    Journal of Systems Science and Mathematical Sciences. 2025, 45(4): 1308-1323. https://doi.org/10.12341/jssms23628
    Abstract & Html ( ) Download PDF ( )   Knowledge map   Save
    In order to improve the shortcomings of the current core indicator estimator in China's sample rotation survey, which does not fully utilize the information of the previous data, this paper investigates the AK composite estimator, which is the international mainstream in sample rotation surveys. It explores the construction method and statistical properties of the AK composite estimator. The main contributions are twofold: First, the AK composite estimator of the 4-8-4 type sample rotation scheme commonly used in existing research is extended to the AK composite estimator of the r-m-r type sample rotation scheme, and its bias, variance and mean square error are given; Second, Bootstrap method is introduced instead of the current balanced half method to construct the sampling variance estimator of the AK composite estimator, and the optimal coefficients of the AK composite estimator are determined by minimizing the mean squared error. Simulation results show that compared with the balanced half method, the Bootstrap method more comprehensively analyzes the sample overlap and sample correlation between the resampled samples in different months, and the optimal weights of the AK composite estimator based on the Bootstrap method have a smaller mean square error. The results of this paper will help optimize and improve China's labor force survey system.
  • GUO Chaohui
    Journal of Systems Science and Mathematical Sciences. 2025, 45(4): 1324-1338. https://doi.org/10.12341/jssms23834
    Abstract & Html ( ) Download PDF ( )   Knowledge map   Save
    Partially linear varying coefficient models have been considered as one of the most popular semiparametric models for prediction because of their high flexibility and good interpretability. However, it is a challenging task to set which covariates have constant or varying coefficient effects when using the partially linear varying coefficient model. In order to effectively reduce the risk of model misspecification caused by artificial setting of constant and varying coefficients, this paper proposes an optimal weighted conditional mean estimation with partially linear varying coefficient model structure based on the idea of model averaging. Specifically, this paper constructs a series of partially linear varying coefficient candidate submodels, each with a different model structure, and obtains the conditional mean estimation under each submodel. Then, the Mallows criterion is used to select model weights to obtain the optimal weighted conditional mean estimation. Under some conditions, this paper proves that the proposed procedure is asymptotically optimal in the sense that its squared prediction risk is asymptotically identical to that of the best but infeasible model averaging estimator though all candidate models are misspecified. Furthermore, derive the rate of the estimated weights converging to the optimal weights minimizing the expected quadratic errors. When the candidate models contain the correct model, this paper shows that our method can put the weight one to the correctly specified models. Finally, simulation examples are used to compare the proposed method with commonly-used model selection and model averaging methods, and the results show that the proposal is superior to existing methods. Furthermore, the data set of selling child car seats at 400 different stores further illustrates that the proposed method is promising.