检索结果-南通市图书馆

Threshold Selection and Resource Allocation for Quantized Identification

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

Journal of systems science & Complexity 2024年第1期37卷 204-229页

作者： WANG Ying LI Xin ZHAO Yanlong ZHANG Ji-Feng Key Laboratory of systems and Control Institute of Systems ScienceAcademy of Mathematics and Systems ScienceChinese Academy of SciencesBeijing 100190China School of Mathematical sciences University of Chinese Academy of SciencesBeijing 100049China

This paper is concerned with the optimal threshold selection and resource allocation problems of quantized identification, whose aims are improving identification efficiency under limited resources. Firstly, the first-order asymptotically optimal quantized identification theory is extended to the weak persistent excitation condition. Secondly, the characteristics of time and space complexities are established based on the Cramér-Rao lower bound of quantized systems. On these basis, the optimal selection methods of fixed thresholds and adaptive thresholds are established under aperiodic signals, which answer how to achieve the best efficiency of quantized identification under the same time and space complexity. In addition, based on the principle of maximizing the identification efficiency under a given resource, the optimal resource allocation methods of quantized identification are given for the cases of fixed thresholds and adaptive thresholds, respectively, which show how to balance time and space complexity to realize the best identification efficiency of quantized identification.

关键词： Quantized output resource allocation system identification threshold selection.

Forward to the Special Topic on“Solving Differential Equations with Deep Learning”

在线全文

学校读者我要写书评

暂无评论

Journal of systems science & Complexity 2024年第2期37卷 389-390页

作者： YAN Zhenya academy of Mathematics and systems science Chinese Academy of SciencesBeijing 100190China

The concept of artificial intelligence(AI)was first presented at the Dartmouth SummerResearch Project on AI hosted by McCarthy and Minsky in 1956.After that,the developmentof AI has experienced ups and downs due to some limited conditions including intelligentcomputing techniques,computer and hardware and software^([1,2]).In particular,with the rapidof deep artificial neural networks,machine learning(ML)and deep learning(DL)techniques^([3,4])have been applied in many fields of science and engineering,including machine translation,image recognition,genomics,scientific computing and atmospheric science,etc^([5]).

关键词： has artificial hardware

A stochastic gradient-based two-step sparse identification algorithm for multivariate ARX systems

在线全文

学校读者我要写书评

暂无评论

Control Theory and Technology 2024年第2期22卷 213-221页

作者： Yanxin Fu Wenxiao Zhao Key Laboratory of systems and Control Academy of Mathematics and Systems ScienceChinese Academy of SciencesBeijing100190China School of Mathematical sciences University of Chinese Academy of SciencesBeijing100049China

We consider the sparse identification of multivariate ARX systems, i.e., to recover the zero elements of the unknown parameter matrix. We propose a two-step algorithm, where in the first step the stochastic gradient (SG) algorithm is applied to obtain initial estimates of the unknown parameter matrix and in the second step an optimization criterion is introduced for the sparse identification of multivariate ARX systems. Under mild conditions, we prove that by minimizing the criterion function, the zero elements of the unknown parameter matrix can be recovered with a finite number of observations. The performance of the algorithm is testified through a simulation example.

关键词： ARX system Stochastic gradient algorithm Sparse identification Support recovery Parameter estimation Strong consistency

Distributed Least Squares Algorithm of Continuous-Time Stochastic Regression Model Based on Sampled Data

维普期刊数据库评论

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Journal of systems science & Complexity 2024年第2期37卷 609-628页

作者： ZHU Xinghua GAN Die LIU Zhixin The Key Laboratory of systems and Control Academy of Mathematics and Systems ScienceChinese Academy of Sciencesand School of Mathematical SciencesUniversity of Chinese Academy of SciencesBeijing 100190China Zhongguancun Laboratory Beijing 100094China

In this paper,the authors consider the distributed adaptive identification problem over sensor networks using sampled data,where the dynamics of each sensor is described by a stochastic differential equation.By minimizing a local objective function at sampling time instants,the authors propose an online distributed least squares algorithm based on sampled data.A cooperative non-persistent excitation condition is introduced,under which the convergence results of the proposed algorithm are established by properly choosing the sampling time interval.The upper bound on the accumulative regret of the adaptive predictor can also be provided.Finally,the authors demonstrate the cooperative effect of multiple sensors in the estimation of unknown parameters by computer simulations.

关键词： Cooperative excitation condition distributed least squares regret sampled data stochastic differential equation

Fiducial Approach for the Storage Reliability Assessment of Complex Repairable systems

在线全文

学校读者我要写书评

暂无评论

Journal of systems science & Complexity 2024年第4期37卷 1653-1671页

作者： YANG Yang ZHAO Letian CHEN Siyi YU Dan University of Chinese academy of sciences Computer Application Institute of Nuclear Industry academy of Mathematics and systems science Chinese Academy of Sciences

In recent years, there has been a growing focus on high-reliability products in industrial and military fields. Storage plays a crucial role in these products, and as a result, research on storage reliability has gained more attention. This paper specifically targets a class of products with complex structures that require long-term storage, regular testing, and maintenance. Firstly, an expression method for system availability is provided based on the reliability structure of the system and the maintenance situation of the constituent equipment in storage. Since the expression of system availability is typically complex and difficult to compute, and the storage life of the system cannot be represented as an explicit function of the reliability indicators of the constituent equipment, it is challenging to evaluate the availability and storage life of repairable systems. To address this, the paper proposes a comprehensive evaluation methodology for assessing the storage reliability of complex repairable systems based on fiducial inference. This approach is employed to derive point estimates and determine the lower fiducial limits for both system availability and storage life. Furthermore, simulation results demonstrate that this comprehensive evaluation method for storage reliability of complex repairable systems is not only convenient but also highly effective.

关键词： Availability fiducial method repairable system storage life

Compensated acceleration feedback based active disturbance rejection control for launch vehicles

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

Chinese Journal of Aeronautics 2024年第4期37卷 464-478页

作者： Xiaoyan ZHANG Wenchao XUE Zibo LIU Ran ZHANG Huifeng LI Key Laboratory of systems and Control Institute of Systems ScienceAcademy of Mathematics and Systems ScienceChinese Academy of SciencesBeijing 100190China School of Mathematical sciences University of Chinese Academy of SciencesBeijing 100049China School of Astronautics Beihang UniversityBeijing 100191China

In this paper, the attitude tracking and load relief control problems against wind disturbances and uncertain aerodynamics as well as the engine thrust of launch vehicles are studied.Firstly, a framework of Compensated Acceleration Feedback based Active Disturbance Rejection Control(CAF-ADRC) is established to achieve both desired attitude tracking and load relief performances. In particular, the total disturbance that includes the effects caused by both aerocoefficient perturbations and disturbances is estimated by constructing an Extended State Observer(ESO) to achieve attitude tracking. Furthermore, combined with the normal acceleration due to the engine thrust, the accelerometer measurement is also compensated to enhance the load relief effect.Secondly, the quantitative analysis of ESO and the entire closed-loop system are studied. It can be concluded that the desired attitude tracking and load relief performances can be achieved simultaneously under the proposed approach. Besides, tuning laws of the proposed approach are systematically given, which are divided into ESO, Proportional Derivative(PD) and Compensated Acceleration Feedback(CAF) modules. Moreover, the performances under CAF-ADRC approach can be better than those under CAF based PD(CAF-PD) approach by tuning load relief gain.Finally, the approach presented is applied to a typical control problem of launch vehicles with wind disturbances and parameter uncertainties.

关键词： Launch vehicles Uncertainty analysis Active disturbance rejection control(ADRC) Load relief control Extended state observer(ESO)

Understanding Human and Machine Interaction from Decision Perspective: An Empirical Study Based on the Game of Go

在线全文

学校读者我要写书评

暂无评论

Journal of systems science & Complexity 2024年第2期37卷 647-667页

作者： ZHAO Ping LI Xuerong WANG Shouyang academy of Mathematics and systems science Chinese Academy of SciencesBeijing 100190China School of Mathematical sciences University of Chinese Academy of SciencesBeijing 100190China

The authors aim to interpret human and AI interactions from the decision perspective.The authors decompose the interaction analysis into the following main components in the context of interactions:Individual behavior patterns,interaction relationships,and comprehensive analysis.The authors interpret intertemporal decisions from a physical perspective and employ cross-discipline concepts and methodologies to extract the behavior characteristics of players in the empirical case study.About the individual behavior patterns,the authors find that human players prefer short-term periods to AI in decision-making.The interaction relationship analysis reveals a dynamic relationship between possible short-term co-movement and nearly counter-movement in the long run.The authors apply principal component analysis to descriptive indicators and discover a regular decision hierarchy.The main behavior pattern of players in the game of Go is switching between careful and daring behaviors.The differences in the decision hierarchies imply a discrepancy of patience between humans and AI.

关键词： Coherence analysis decision hierarchy machine behaviour multitaper spectral analysis

Enhancing cut selection through reinforcement learning

在线全文

学校读者我要写书评

暂无评论

science China Mathematics 2024年第6期67卷 1377-1394页

作者： Shengchao Wang Liang Chen Lingfeng Niu Yu-Hong Dai academy of Mathematics and systems science Chinese Academy of SciencesBeijing100190China Research Center on Fictitious Economy and Data science Chinese Academy of SciencesBeijing100190China

With the rapid development of artificial intelligence in recent years,applying various learning techniques to solve mixed-integer linear programming(MILP)problems has emerged as a burgeoning research domain.Apart from constructing end-to-end models directly,integrating learning approaches with some modules in the traditional methods for solving MILPs is also a promising direction.The cutting plane method is one of the fundamental algorithms used in modern MILP solvers,and the selection of appropriate cuts from the candidate cuts subset is crucial for enhancing efficiency.Due to the reliance on expert knowledge and problem-specific heuristics,classical cut selection methods are not always transferable and often limit the scalability and generalizability of the cutting plane method.To provide a more efficient and generalizable strategy,we propose a reinforcement learning(RL)framework to enhance cut selection in the solving process of MILPs.Firstly,we design feature vectors to incorporate the inherent properties of MILP and computational information from the solver and represent MILP instances as bipartite graphs.Secondly,we choose the weighted metrics to approximate the proximity of feasible solutions to the convex hull and utilize the learning method to determine the weights assigned to each metric.Thirdly,a graph convolutional neural network is adopted with a self-attention mechanism to predict the value of weighting factors.Finally,we transform the cut selection process into a Markov decision process and utilize RL method to train the model.Extensive experiments are conducted based on a leading open-source MILP solver SCIP.Results on both general and specific datasets validate the effectiveness and efficiency of our proposed approach.

关键词： reinforcement learning mixed-integer linear programming cutting plane method cut selection

Density peak clustering using tensor network

在线全文

学校读者我要写书评

暂无评论

science China(Information sciences) 2024年第3期67卷 321-322页

作者： Xiao SHI Yun SHANG Institute of Mathematics Academy of Mathematics and Systems ScienceChinese Academy of Sciences School of Mathematical sciences University of Chinese Academy of Sciences National Center for Mathematics and Interdisciplinary sciences Key Laboratory of ManagementDecision and Information SystemsAcademy of Mathematics and Systems ScienceChinese Academy of Sciences

Tensor networks have been a powerful tool in simulating many-body physics and have recently gained recognition in the machine learning community due to their remarkable representation capabilities. However, using tens... 详细信息

关键词： Tensors