神经网络求解TSP问题的理论分析及其改进

经过对 TSP(Travelling Salesman Problem)网络的较为完整的特征分析,证明了只有在网络运行的初始状态有效解方向和无效解方向才是网络的特征方向,可以通过控制网络连接权矩阵的特征值的数值来控制这时网络向有效解方向运行,而在网络的运行过程中却无法控制网络向有效解方向运行.只能保证网络向能量下降方向运行,这是网络出现无效解的根本原因,尽管如此,文中通过改进网络的能量函数,从而改进网络的特征值来控制网络的初始运行行为,实验结果表明,改进后的网络具有有效解空间大、收敛速度快、易于获得满意解等优点.

欠自由度机器人的运动学逆解存在性神经网络判别

针对欠自由度机器人在工作空间内运动学逆解的存在性问题 ,提出虚拟关节法 ,并利用神经网络的学习能力 ,区分出可解子空间和不可解子空间 。

线性方程组通解的一种求法

一般线性方程组有无穷多解时,通常先要求出相应的齐次线性方程组,即它的导出组的基础解系,再将一般线性方程组的通解表示为它的一个特解与导出组基础解系的和的形式.通过引进增广齐次方程组和它的条件解的概念,给出了由求增广齐次线性方程在xn+1=1下的条件解,同步求出一般线性方程组通解的方法,并且推出了相应的表示一般线性方程组无穷解集的简明表达式,即增广齐次线性方程的一个条件解与Fn+1中的n-r维解子空间和的形式.

The Interpolation of Operators between Weighted Hardy Spaces and Weighted L^p Spaces

The paper is given the interpolation of operators between weighted Hardy spaces and weighted L p spaces when w∈A 1 by Calderon Zygmund decomposition.

Iterative Solution for Systems of Binary Non-mixed Monotone Operator Equations

With using the cone and partial ordering t heory and mixed monotone operator theory, the existence and uniqueness for solut ion of systems of non-monotone binary nonliear operator equations are discussed. And the iterative sequences which converge to solution of systems of operator e quations and the error estimates are also given. Some corresponding results for the mixed monotone operations and the unary operator equations are improved and generalized.

Multipliers on Weighted Hardy Spaces over Locally Compact Vilenkin Groups

This paper is devoted to study the atomic decomposition of the weighted Hardy space on Vilenkin groups and a weighted Hp multiplier theorem.

Constrained least squares algorithm for channel vector estimation in 2-D RAKE receiver

Based on the fact that the variation of tile direction of arrival (DOA) isslower than that of the channel fading, the steering vector of the desired signal is estimatedfirstly using a subspace decomposition method and then a constrained condition is configured.Traffic signals are further employed to estimate the channel vector based on the constrained leastsquares criterion. We use the iterative least squares with projection (ILSP) algorithm initializedby the pilot to get the estimation. The accuracy of channel estimation and symbol detection can beprogressively increased through the iteration procedure of the ILSP algorithm. Simulation resultsdemonstrate that the proposed algorithm improves the system performance effectively compared withthe conventional 2-D RAKE receiver.

On Weighted Tent Space

In this paper, we introduce the A, weights into the tent space, many important results in the tent space are generalized. Also, new relations between the A, weights and Carleson measures are obtained.

Anti-periodic solutions to a class of second-order evolution equations

In this paper we discuss the anti-periodic problem for a class of abstractnonlinear second-order evolution equations associated with maximal monotone operators in Hilbertspaces and give some new assumptions on operators. We establish the existence and uniqueness ofanti-periodic solutions, which improve andgeneralize the results that have been obtained. Finally weillustrate the abstract theory by discussing a simple example of an anti-periodic problem fornonlinear partial differential equations.

Classification of hyperspectral remote sensing images based on simulated annealing genetic algorithm and multiple instance learning

A hybrid feature selection and classification strategy was proposed based on the simulated annealing genetic algonthrn and multiple instance learning （MIL）. The band selection method was proposed from subspace decomposition, which combines the simulated annealing algorithm with the genetic algorithm in choosing different cross-over and mutation probabilities, as well as mutation individuals. Then MIL was combined with image segmentation, clustering and support vector machine algorithms to classify hyperspectral image. The experimental results show that this proposed method can get high classification accuracy of 93.13% at small training samples and the weaknesses of the conventional methods are overcome.

Rational design of stratified material with spatially separated catalytic sites as an efficient overall water-splitting photocatalyst

The development of metal sulfide catalysts with remarkable activity toward efficient overall photocatalytic water splitting remains challenging owing to the dominant charge recombination and deficient catalytic active sites.Moreover,in the process of water oxidation catalysis,the inhibition of severe photocorrosion is an immense task,requiring effective photogenic hole-transfer kinetics.Herein,stratified Co-MnO_(2)@CdS/CoS hollow cubes with spatially separated catalytic sites were rationally designed and fabricated as highly efficient controllable catalysts for photocatalytic overall water splitting.The unique self-templated method,including a continuous anion/cation-exchange reaction,integrates a Co-doped oxidation co-catalyst(Co-MnO_(2))and a reduction co-catalyst(CoS)on the nanocubes with uniform interface contact and ultrathin two-dimensional(2D)nanometer sheets.We demonstrate that the stratified Co-MnO_(2)@CdS/CoS hollow cubes can provide an abundance of active sites for surface redox reactions and contribute to the separation and migration of the photoionization charge carriers.In particular,CoS nanoparticles dispersed on the walls of CdS hollow cubes were identified as reduction co-catalysts accelerating hydrogen generation,while Co-MnO_(2) nanosheets attached to the inner walls of the CdS hollow cube were oxidation co-catalysts,promoting oxygen evolution dynamics.Benefiting from the desirable structural and compositional advantages,optimized stratification of Co-MnO_(2)@CdS/CoS nanocubes provided a catalytic system devoid of precious metals,which exhibited a remarkable overall photocatalytic water-splitting rate(735.4(H_(2))and 361.1(O_(2))μmol h^(−1) g^(−1)),being among the highest values reported thus far for CdS-based catalysts.Moreover,an apparent quantum efficiency(AQE)of 1.32%was achieved for hydrogen evolution at 420 nm.This study emphasizes the importance of rational design on the structure and composition of photocatalysts for overall water splitting.

A Remark on Concircularly Flat Totally Real Minimal Submanifolds in CP^n

Let M be a concircularly fiat totally real minimal submanifold in CP4. The infimum Vm of the volume V （M） of M is obtained, also the necessary and sufficient conditions of ＂V（M）=Vm＂ is given.

Unsupervised linear spectral mixture analysis with AVIRIS data

A new algorithm for unsupervised hyperspectral data unmixing is investigated, which includes a modified minimum noise fraction （MNF） transformation and independent component analysis （ICA）. The modified MNF transformation is used to reduce noise and remove correlation between neighboring bands. Then the ICA is applied to unmix hyperspectral images, and independent endmembers are obtained from unmixed images by using post-processing which includes image segmentation based on statistical histograms and morphological operations. The experimental results demonstrate that this algorithm can identify endmembers resident in mixed pixels. Meanwhile, the results show the high computational efficiency of the modified MNF transformation. The time consumed by the modified method is almost one fifth of the traditional MNF transformation.

Multipliers on Generalized Bergman Spaces

In this paper,we characterize the multipliers of generalized Bergman spaces A^p，q，α with 0＜p≤1, 0＜q, α<∞into some analytic function spaces and into sequence spaces,and show that the multipliers of A^p，q，α(0<p≤1,0<q, α<∞) into a given space are the same as those of A^p，α(0<p≤1, α>0) in almost every case considered. The corollaries on multipliers of the spaces A^p，q，α extend some related results.

NEW FAST ALGORITHM FOR 2-D ANGLE-OF-ARRIVAL ESTIMATION

Based on propagator method, a fast 2-D Angle-Of-Arrival (AOA) algorithm is proPosed in this paper. The proposed algorithm does not need the Eigen-Value Decomposition (EVD) or Singular Value Decomposition (SVD) of the Sample Covariance Matrix (SCM), thus the fast algorithm has lower computational complexity with insignificant performance degradation when comparing with conventional subspace approaches. Furthermore, the proposed algorithm has no performance degradation. Finally, computer simulations verify the effectiveness of the proposed algorithm.

Exact Solution of D-Dimensional Klein-Gordon Oscillator with Minimal Length

Specific modifications of the usual canonical commutation relations between position and momentumoperators have been proposed in the literature in order to implement the idea of the existence of a minimal observablelength. Here, we study a consequence of this proposal in relativistic quantum mechanics by solving in the momentumspace representation the Klein-Gordon oscillator in arbitrary dimensions. The exact bound states spectrum and thecorresponding momentum space wave function are obtained. Following Chang et al. (Phys. Rev. D 65 (2002) 125027),we discuss constraint that can be placed on the minimal length by measuring the energy levels of an electron in a penningtrap.

Atomic decomposition of Besov-type and Triebel-Lizorkin-type spaces

With the help of the maximal function caracterizations of the Besov-type space Bs,τ p,q and the Triebel- Lizorkin-type space Fs,τ p,q, we present the atomic decomposition of these function spaces. Our results cover the results on classical Besov and Triebel-Lizorkin spaces by taking T = 0.

Generalized Morrey spaces and trace operator

The theory of generalized Besov-Morrey spaces and generalized Triebel-Lizorkin-Morrey spaces is developed. Generalized Morrey spaces, which Mizuhara and Nakai proposed, are equipped with a parameter and a function. The trace property is one of the main focuses of the present paper, which will clarify the role of the parameter of generalized Morrey spaces. The quarkonial decomposition is obtained as an application of the atomic decomposition. In the end, the relation between the function spaces dealt in the present paper and the foregoing researches is discussed.

B-valued martingale spaces with small index and atomic decompositions

Some atomic decomposition theorems for Banach-space-valued martingales are established. Using them, the embedding relationships between martingale spaces with small index are discussed. The results obtained here are connected closely with the p-uniform smoothness and q-uniform convexity of Banach space in which the martingales take values.

Atomic decompositions and Hardy's inequality on weak Hardy-Morrey spaces

We introduce the weak Hardy-Morrey spaces in this paper.We also obtain the atomic decompositions of the weak Hardy-Morrey spaces.By using these decompositions,we establish the Hardy inequalities on the weak Hardy-Morrey spaces.