针对深度Q网络(DQN)算法因过估计导致收敛稳定性差的问题,在传统时序差分(TD)的基础上提出N阶TD误差的概念,设计基于二阶TD误差的双网络DQN算法。构造基于二阶TD误差的值函数更新公式,同时结合DQN算法建立双网络模型,得到两个同构的值...针对深度Q网络(DQN)算法因过估计导致收敛稳定性差的问题,在传统时序差分(TD)的基础上提出N阶TD误差的概念,设计基于二阶TD误差的双网络DQN算法。构造基于二阶TD误差的值函数更新公式,同时结合DQN算法建立双网络模型,得到两个同构的值函数网络分别用于表示先后两轮的值函数,协同更新网络参数,以提高DQN算法中值函数估计的稳定性。基于Open AI Gym平台的实验结果表明,在解决Mountain Car和Cart Pole问题方面,该算法较经典DQN算法具有更好的收敛稳定性。展开更多
The main aim of this paper is to have an accurate analysis on the famous Adini＇s element for the second order problems under to the anisotropic meshes. We firstly show that the interpolation of Adini＇s element satis...The main aim of this paper is to have an accurate analysis on the famous Adini＇s element for the second order problems under to the anisotropic meshes. We firstly show that the interpolation of Adini＇s element satisfy the anisotropic property. Then the optimal error estimate is obtained without the regularity assumption on the meshes.展开更多
To improve the deteriorated capacity gain and source recovery performance due to channel mismatch problem,this paper reports a research about blind separation method against channel mismatch in multiple-input multiple...To improve the deteriorated capacity gain and source recovery performance due to channel mismatch problem,this paper reports a research about blind separation method against channel mismatch in multiple-input multiple-output(MIMO) systems.The channel mismatch problem can be described as a channel with bounded fluctuant errors due to channel distortion or channel estimation errors.The problem of blind signal separation/extraction with channel mismatch is formulated as a cost function of blind source separation(BSS) subject to the second-order cone constraint,which can be called as second-order cone programing optimization problem.Then the resulting cost function is solved by approximate negentropy maximization using quasi-Newton iterative methods for blind separation/extraction source signals.Theoretical analysis demonstrates that the proposed algorithm has low computational complexity and improved performance advantages.Simulation results verify that the capacity gain and bit error rate(BER) performance of the proposed blind separation method is superior to those of the existing methods in MIMO systems with channel mismatch problem.展开更多
This work is concerned with time stepping finite element methods for abstract second order evolution problems. We derive optimal order a posteriori error estimates and a posteriori nodal superconvergence error estimat...This work is concerned with time stepping finite element methods for abstract second order evolution problems. We derive optimal order a posteriori error estimates and a posteriori nodal superconvergence error estimates using the energy approach and the duality argument. With the help of the a posteriori error estimator developed in this work, we will further propose an adaptive time stepping strategy. A number of numerical experiments are performed to illustrate the reliability and efficiency of the a posteriori error estimates and to assess the effectiveness of the proposed adaptive time stepping method.展开更多
In this paper, we report our recent advances on vertex centered finite volume element methods （FVEMs） for second order partial differential equations （PDEs）. We begin with a brief review on linear and quadratic fi...In this paper, we report our recent advances on vertex centered finite volume element methods （FVEMs） for second order partial differential equations （PDEs）. We begin with a brief review on linear and quadratic finite volume schemes. Then we present our recent advances on finite volume schemes of arbitrary order. For each scheme, we first explain its construction and then perform its error analysis under both HI and L2 norms along with study of superconvergence properties.展开更多
The cross-coupled control(CCC)is widely applied to reduce contour errors in contour-following applications.In such situation,the contour error estimation plays an important role.Traditionally,the linear or second-orde...The cross-coupled control(CCC)is widely applied to reduce contour errors in contour-following applications.In such situation,the contour error estimation plays an important role.Traditionally,the linear or second-order estimation approach is adopted for biaxial motion systems,whereas only linear approach is available for triaxial systems.In this paper,the second-order contour error estimation,which was presented in our previous work,is utilized to determine the variable CCC gains for motion control systems with three axes.An integrated stable motion control strategy,which combines the feedforward,feedback and CCC controllers,is developed for multiaxis CNC systems.Experimental results on a triaxial platform indicate that the CCC scheme based on the second-order estimation,compared with that based on the linear one,significantly reduces the contour error even in the conditions of high tracking feedrate and small radius of curvature.展开更多
文摘针对深度Q网络(DQN)算法因过估计导致收敛稳定性差的问题,在传统时序差分(TD)的基础上提出N阶TD误差的概念,设计基于二阶TD误差的双网络DQN算法。构造基于二阶TD误差的值函数更新公式,同时结合DQN算法建立双网络模型,得到两个同构的值函数网络分别用于表示先后两轮的值函数,协同更新网络参数,以提高DQN算法中值函数估计的稳定性。基于Open AI Gym平台的实验结果表明,在解决Mountain Car和Cart Pole问题方面,该算法较经典DQN算法具有更好的收敛稳定性。
基金the Henan Natural Science Foundation(072300410320)the Henan Education Department Foundational Study Foundation(200510460311)
文摘The main aim of this paper is to have an accurate analysis on the famous Adini＇s element for the second order problems under to the anisotropic meshes. We firstly show that the interpolation of Adini＇s element satisfy the anisotropic property. Then the optimal error estimate is obtained without the regularity assumption on the meshes.
基金supported by Sichuan Youth Science and Technology Innovation Research Team Project(No.2015TD0022)the Talents Project of Sichuan University of Science and Engineering(No.2017RCL11 and No.2017RCL10)the first batch of science and technology plan key R&D project of Sichuan province(No.2017GZ0068)
文摘To improve the deteriorated capacity gain and source recovery performance due to channel mismatch problem,this paper reports a research about blind separation method against channel mismatch in multiple-input multiple-output(MIMO) systems.The channel mismatch problem can be described as a channel with bounded fluctuant errors due to channel distortion or channel estimation errors.The problem of blind signal separation/extraction with channel mismatch is formulated as a cost function of blind source separation(BSS) subject to the second-order cone constraint,which can be called as second-order cone programing optimization problem.Then the resulting cost function is solved by approximate negentropy maximization using quasi-Newton iterative methods for blind separation/extraction source signals.Theoretical analysis demonstrates that the proposed algorithm has low computational complexity and improved performance advantages.Simulation results verify that the capacity gain and bit error rate(BER) performance of the proposed blind separation method is superior to those of the existing methods in MIMO systems with channel mismatch problem.
基金supported by National Natural Science Foundation of China(Grant Nos.1117121911161130004 and 11101199)+1 种基金E-Institutes of Shanghai Municipal Education Commission(Grant No.E03004)Program for New Century Excellent Talents in Fujian Province University(Grant No.JA12260)
文摘This work is concerned with time stepping finite element methods for abstract second order evolution problems. We derive optimal order a posteriori error estimates and a posteriori nodal superconvergence error estimates using the energy approach and the duality argument. With the help of the a posteriori error estimator developed in this work, we will further propose an adaptive time stepping strategy. A number of numerical experiments are performed to illustrate the reliability and efficiency of the a posteriori error estimates and to assess the effectiveness of the proposed adaptive time stepping method.
基金supported by National Science Foundation of USA(Grant No.DMS1115530)National Natural Science Foundation of China(Grant No.11171359)the Fundamental Research Funds for the Central Universities of China
文摘In this paper, we report our recent advances on vertex centered finite volume element methods （FVEMs） for second order partial differential equations （PDEs）. We begin with a brief review on linear and quadratic finite volume schemes. Then we present our recent advances on finite volume schemes of arbitrary order. For each scheme, we first explain its construction and then perform its error analysis under both HI and L2 norms along with study of superconvergence properties.
基金supported by the National Natural Science Foundation of China(Grant Nos.51325502 and 51405175)the National Basic Research Program of China("973"Project)(Grant No.2011CB706804)the National Science and Technology Major Projects of China(Grant No.2012ZX04001-012-01-05)
文摘The cross-coupled control(CCC)is widely applied to reduce contour errors in contour-following applications.In such situation,the contour error estimation plays an important role.Traditionally,the linear or second-order estimation approach is adopted for biaxial motion systems,whereas only linear approach is available for triaxial systems.In this paper,the second-order contour error estimation,which was presented in our previous work,is utilized to determine the variable CCC gains for motion control systems with three axes.An integrated stable motion control strategy,which combines the feedforward,feedback and CCC controllers,is developed for multiaxis CNC systems.Experimental results on a triaxial platform indicate that the CCC scheme based on the second-order estimation,compared with that based on the linear one,significantly reduces the contour error even in the conditions of high tracking feedrate and small radius of curvature.
