首页期刊导航|中国科学:信息科学(英文版)
期刊信息/Journal information
中国科学:信息科学(英文版)
中国科学:信息科学(英文版)

周光召

月刊

1674-733X

informatics@scichina.org

010-64015683

100717

北京东黄城根北街16号

中国科学:信息科学(英文版)/Journal Science China Information SciencesCSCDCSTPCDEISCI
查看更多>>《中国科学》是中国科学院主办、中国科学杂志社出版的自然科学专业性学术刊物。《中国科学》任务是反映中国自然科学各学科中的最新科研成果,以促进国内外的学术交流。《中国科学》以论文形式报道中国基础研究和应用研究方面具有创造性的、高水平的和有重要意义的科研成果。在国际学术界,《中国科学》作为代表中国最高水平的学术刊物也受到高度重视。国际上最具有权威的检索刊物SCI,多年来一直收录《中国科学》的论文。1999年《中国科学》夺得国家期刊奖的第一名。
正式出版
收录年代

    An efficient binary programming method for black-box optimization and its application in processor design

    Xiaoliang LVQiaozhu ZHAIJianchen HUYuhang ZHU...
    136-152页
    查看更多>>摘要:Optimizing the parameter settings in a large design space for the processor with limited simula-tion resources is a challenging task.The current black-box optimization algorithms for processor design space exploration(DSE)problems usually require a large amount of simulation resources for high-dimensional and discrete problems.Besides,the constraints handling techniques in these algorithms need to be improved.To address the issues,we propose an efficient binary integer programming(BIP)approach for the DSE of the processor with strictly guaranteed constraints.Our approach involves adopting the separability assumption to establish a surrogate objective function that is ordinal consistent,thus avoiding the complex non-linearity of the real objective function.Moreover,the design rules can be taken simply as constraints in BIP model to further reduce the design space.Thus,the efforts spent in the infeasible exploration space can be avoided.The experimental results show that the proposed algorithm outperforms the state-of-the-art Bayesian op-timization and evolutionary algorithms in terms of exploration efficiency,required simulation points and performance of the recommended points.

    A novel memetic algorithm for distributed shape formation of swarm robots with both acceleration and velocity constraints

    Yun QUBin XINQinqin WANGRuocheng LI...
    153-171页
    查看更多>>摘要:This article investigates the problem of distributed shape formation(DSF)in swarm robot systems.DSF involves each robot autonomously selecting and moving toward a target point to achieve a desired shape.A comprehensive mathematical model for DSF of swarm robots is developed,capturing the relationships among states,behaviors,and kinematics constraints.A novel memetic algorithm(MA)is proposed to generate and evolve behavior strategies that enable real-time decision-making for each robot.The proposed MA includes specifically designed behaviors to address subproblems within DSF,such as target selection conflict,collision,and deadlock.These behaviors,along with propositions about the states of robots,are utilized to construct strategies by a tree-based encoding scheme.An evaluation mechanism based on multi-agent simulation is devised to evaluate the performance of strategies.Additionally,a repair mechanism is introduced to eliminate redundant or unreachable subtrees in a strategy.To further improve the performance of generated strategies,tree-based local search operators are employed to exploit the neighborhood of the best strategy found yet during the iteration.Experimental results and the Wilcoxon rank-sum test show that the strategies generated by the proposed MA outperform state-of-the-art algorithms,significantly reducing the completion time of DSF.

    Graph-geometric message passing via a graph convolution transformer for FKP regression

    Huizhi ZHUWenxia XUJian HUANGBaocheng YU...
    172-186页
    查看更多>>摘要:In this paper,the forward kinematics problem(FKP)of the Gough-Stewart platform(GSP)with six degrees of freedom(6 DoFs)is estimated via deep learning.We propose a graph convolution transformer model by systematically analyzing some challenges encountered with using deep learning regression on large-scale data.We attempt to leverage the graph-geometric message as input and singular value decomposition(SVD)orthogonalization for SO(3)manifold learning.This study is the first in which a robot with a sophisticated closed-loop mechanism is described by a graph structure and a specific deep learning model is proposed to solve the FKP of the GSP.Qualitative and quantitative experiments on our dataset demonstrate that our model is feasible and superior to other methods.Our method can guarantee error percentages of translation and rotation less than 1 mm and 1° of 81.9%and 96.7%,respectively.

    Time-varying formation tracking control of high-order multi-agent systems with multiple leaders and multiplicative noise

    Ruru JIAXiaofeng ZONGQing WANG
    187-203页
    查看更多>>摘要:This work discusses the time-varying formation(TVF)tracking control problem of high-order multi-agent systems(MASs)with multiple leaders and multiplicative measurement noise.With the help of Lyapunov function tools and stochastic analysis methods,the TVF tracking protocol with multiple leaders and multiplicative noise is developed based on the relative state measurements,where followers are driven to realize the target TVF while tracking the convex combination formed by multiple leaders.Then,the TVF tracking problem is converted into the mean square asymptotic stability problem of a stochastic differential equation(SDE);sufficient conditions related to the control gains are given by stabilizing the corresponding stochastic system.Moreover,a TVF tracking algorithm is presented to outline the steps of protocol design.Finally,the theoretical results are illustrated in terms of simulation examples.

    Controllability of descriptor multi-agent systems with signed networks

    Yu SHENYongqiang GUANYe TIAN
    204-222页
    查看更多>>摘要:This paper studies the controllability of descriptor multi-agent systems(DMASs)with signed networks,where the networks are signed,and agents have descriptor linear dynamics.First,the basis for the eigenspace of the system matrix is derived completely and qualitatively by leveraging the maximum general-ized left Jordan chain defined in this paper.Taking advantage of the explicit form of the basis,necessary and sufficient conditions for the controllability of DMASs are established.Furthermore,a necessary and sufficient condition is provided to ensure the controllability of DMASs with heterogeneous dynamics.Particularly,for some special cases,controllability conditions are expressed more precisely in terms of eigenvectors.Then,the relationship between the controllability of DMASs with signed and unsigned networks is investigated.It is shown that the controllability of DMASs under structurally balanced signed networks is equivalent to that under the associated underlying unsigned networks whether the individual dynamics is homogeneous or heterogeneous.Finally,theoretical results are applied to the multi-agent supporting systems.

    A distributed decomposition algorithm for solving large-scale mixed integer programming problem

    Fangzheng TIANHongzhe LIUWenwu YU
    223-235页
    查看更多>>摘要:Mixed integer programming is inherently involved in solving a significant number of practical problems.This paper focuses on mixed integer programming,where the objective function is the summation of N functions,and the constraints include both scalar coupling and set constraints.Given the potentially large scale of these problems,the goal of this work is to propose a distributed method to solve large-scale problems more efficiently.The right-hand side allocation decomposition approach is employed to address the large-scale mixed integer programming problem.Algorithms are then proposed for solving these problems,based on the analysis of the continuity,differentiability,and local convexity properties of the decomposed subproblems.Simulation experiments with randomly generated coefficients demonstrate the superior perfor-mance of the proposed algorithms compared to the Gurobi solver,offering higher solution accuracy and faster processing time for large-scale mixed integer programming problems with nonlinear objective and constraint functions.

    Learning continuous network emerging dynamics from scarce observations via data-adaptive stochastic processes

    Jiaxu CUIQipeng WANGBingyi SUNJiming LIU...
    236-251页
    查看更多>>摘要:Learning network dynamics from the empirical structure and spatio-temporal observation data is crucial to revealing the interaction mechanisms of complex networks in a wide range of domains.However,most existing methods only aim at learning network dynamic behaviors generated by a specific ordinary dif-ferential equation instance,resulting in ineffectiveness for new ones,and generally require dense observations.The observed data,especially from network emerging dynamics,are usually difficult to obtain,which brings trouble to model learning.Therefore,learning accurate network dynamics with sparse,irregularly-sampled,partial,and noisy observations remains a fundamental challenge.We introduce a new concept of the stochas-tic skeleton and its neural implementation,i.e.,neural ODE processes for network dynamics(NDP4ND),a new class of stochastic processes governed by stochastic data-adaptive network dynamics,to overcome the challenge and learn continuous network dynamics from scarce observations.Intensive experiments conducted on various network dynamics in ecological population evolution,phototaxis movement,brain activity,epi-demic spreading,and real-world empirical systems,demonstrate that the proposed method has excellent data adaptability and computational efficiency,and can adapt to unseen network emerging dynamics,producing accurate interpolation and extrapolation with reducing the ratio of required observation data to only about 6%and improving the learning speed for new dynamics by three orders of magnitude.

    Neural Liénard system:learning periodic manipulation skills through dynamical systems

    Haoyu ZHANGLong CHENGYu ZHANGYifan WANG...
    252-267页
    查看更多>>摘要:Learning from demonstrations provides effective methods for teaching robot manipulation skills.However,capturing periodic manipulation skills remains challenging with the current techniques.To address this gap,we introduce the neural Liénard system(neural LS),a novel neural network framework that utilizes Liénard-type differential equations to create dynamical systems with stable and distinct limit cycles.We also introduce an innovative technique,which not only manages periodic trajectories across various dimensions but also handles trajectories with intersections,enhancing the capability of the robot for complex tasks.We provide a thorough theoretical analysis of neural LS,focusing on its stability and representational capabilities.Empirical evaluations show that neural LS achieves superior performance in modeling complex limit cycles,surpassing the existing methods.We particularly emphasize its effectiveness in handling high-dimensional periodic motions and trajectories with intersections.In addition,we explore the adaptability and robustness of neural LS.A practical application involving the Franka Emika robot in a drawing task further demonstrates the real-world utility of neural LS,confirming its effectiveness and potential to equip robots with advanced periodic manipulation skills.

    Optimization methods rooted in optimal control

    Huanshui ZHANGHongxia WANGYeming XUZiyuan GUO...
    268-276页
    查看更多>>摘要:In the paper,we investigate the optimization problem(OP)by applying the optimal control method.The optimization problem is reformulated as an optimal control problem(OCP)where the con-troller(iteration updating)is designed to minimize the sum of costs in the future time instant,which thus theoretically generates the"optimal algorithm"(fastest and most stable).By adopting the maximum prin-ciple and linearization with Taylor expansion,new algorithms are proposed.It is shown that the proposed algorithms have a superlinear convergence rate and thus converge more rapidly than the gradient descent;meanwhile,they are superior to Newton's method because they are not divergent in general and can be ap-plied in the case of a singular or indefinite Hessian matrix.More importantly,the OCP method contains the gradient descent and the Newton's method as special cases,which discovers the theoretical basis of gradient descent and Newton's method and reveals how far these algorithms are from the optimal algorithm.The merits of the proposed optimization algorithm are illustrated by numerical experiments.

    Ultra-low power IGZO optoelectronic synaptic transistors for neuromorphic computing

    Li ZHUSixian LIJunchen LINYuanfeng ZHAO...
    277-286页
    查看更多>>摘要:Inspired by biological visual systems,optoelectronic synapses with image perception,memory retention,and preprocessing capabilities offer a promising pathway for developing high-performance artificial perceptual vision computing systems.Among these,oxide-based optoelectronic synaptic transistors are well-known for their enduring photoconductive properties and ease of integration,which hold substantial potential in this regard.In this study,we utilized indium gallium zinc oxide as a semiconductor layer and high-k Zr AlOx as a gate dielectric layer to engineer low-power high-performance synaptic transistors with photonic memory.Crucial biological synaptic functions,including excitatory postsynaptic currents,paired-pulse facilitation,and the transition from short-term to long-term plasticity,were replicated via optical pulse modulation.This simulation was sustained even at an operating voltage as low as 0.0001 V,exhibiting a conspicuous photonic synaptic response with energy consumption as low as 0.0845 fJ per synaptic event.Furthermore,an optoelectronic synaptic device was employed to model"learn-forget-relearn"behavior similar to that exhibited by the human brain,as well as Morse code encoding.Finally,a 3 × 3 device array was constructed to demonstrate its advantages in image recognition and storage.This study provides an effective strategy for developing readily integrable,ultralow-power optoelectronic synapses with substantial potential in the domains of morphological visual systems,biomimetic robotics,and artificial intelligence.