矩阵方程AX=B的Hermitian-广义Hamiltonian矩阵解的迭代算法

利用正交投影、Hermitian-广义Hamiltonian矩阵类的结构与性质及奇异值分解,讨论了矩阵方程AX=B的Hermitian-广义Hamiltonian矩阵解及其最佳逼近的迭代算法,证明了算法的收敛性,求出了相应的最佳逼近解,并给出了相应的算法步骤和数值例子.     

一类非线性波动方程局部解的存在性

通过对一类具非线性阻尼项的波动方程的初边值问题的研究,在非线性阻尼项及源项的指数不加限制的情况下,应用压缩映射原理,结合内插定理和嵌入定理,证明了该问题局部广义解的存在性和唯一性.     

矩阵方程AXB-CXD=R的多项式解法

将利用线性变化,构造一多项式,从而将矩阵方程AXB-CXD=R转化为一容易求解的方程,并给出了矩阵方程AXB-CXD=R有唯一解时的显示表达式X=-(Ck+1)-1Sk(R)E-1或X=F-1Sk(R)(Bk+1)-1,所得到的结果推广了有关文献的相关结论.     

多元函数极值的正定矩阵判定定理的推广

利用矩阵的分块及正定矩阵来处理特殊多元函数的极值问题,通过降低变量元维数的方法,使求三元或三元以上的特殊多元函数的极值成为切实可行.     

仿射非线性系统状态方程的任意阶近似解

针对典型的仿射非线性系统,采用常微分方程理论对其进行求解.首先将系统在平衡点附近进行展开,求得其齐次方程的解,然后利用常数变易法将非线性微分方程变为等价的第二类非线性Volterra积分方程.采用逐次逼近法,求得任意阶近似解,并证明解的收敛性.     

(3+1)维孤子方程的周期孤波解

扩展了Hirota法,即将Hirota法中的测试函数用新的测试函数来替代,并利用扩展了的方法来构造(3+1)维孤子方程的新的周期孤波解、周期双孤波解、双周期双孤波解.显然扩展的Hirota方法也可以解其他一些非线性发展方程.     

变量矩阵算法及一类曲线最佳拟合问题的解

讨论1种对无约束目标函数采用变量矩阵算法求解最优值的方法,变量矩阵算法的特点是利用矩阵迭代计算,收敛速度快,过程较稳定.对于一类可逼近模拟曲线的最佳拟合,可以迭代出符合要求的参数值.作为较典型的应用实例。设计1种幅度均衡器.通过迭代搜索找出满足衰耗误差的元件值,说明曲线拟合达到了预期的效果.     

关于高维波动方程的Cauchy问题的解

Ｐｏｉｓｏｎ公式给出了高维波动方程的Ｃａｕｃｈｙ问题的解,在三维情况下,解的形式为一曲面积分;在二维情况下,解的形式为二重积分．本文证明,当初值条件φ,ψ及自由项ｆ满足一定条件时,Ｃａｕｃｈｙ问题的解可以以偏导数的形式给出．     

关于Diophantine方程y2=px(x2+2)的一点注记

设p是奇素数,运用初等数论方法证明了:如果P=16k4+1,这里k为正奇数,则方程y2=px(x2+2)无正整数解(x,y).     

Multigrid solution of the nonlinear Poisson-Boltzmann equation and calculation of titration curves

Although knowledge of the pKa values and charge states of individual residues is critical to understanding the role of electrostatic effects in protein structure and function, calculating these quantities is challenging because of the sensitivity of these parameters to the position and distribution of charges. Values for many different proteins which agree well with experimental results have been obtained with modified Tanford-Kirkwood theory in which the protein is modeled as a sphere (reviewed in Ref. 1); however, convergence is more difficult to achieve with finite difference methods, in which the protein is mapped onto a grid and derivatives of the potential function are calculated as differences between the values of the function at grid points (reviewed in Ref. 6). Multigrid methods, in which the size of the grid is varied from fine to coarse in several cycles, decrease computational time, increase rates of convergence, and improve agreement with experiment. Both the accuracy and computational advantage of the multigrid approach increase with grid size, because the time required to achieve a solution increases slowly with grid size. We have implemented a multigrid procedure for solving the nonlinear Poisson-Boltzmann equation, and, using lysozyme as a test case, compared calculations for several crystal forms, different refinement procedures, and different charge assignment schemes. The root mean square difference between calculated and experimental pKa values for the crystal structure which yields best agreement with experiment (1LZT) is 1.1 pH units, with the differences in calculated and experimental pK values being less than 0.6 pH units for 16 out of 21 residues. The calculated titration curves of several residues are biphasic.