Backward error analysis of a class of the periodic symplectic matrices for eigenproblem 一類周期辛矩陣對特征值問題的向后誤差分析
They have a number of desirable properties not possessed by wavelets of daubechies type , namely : they have symmetry property ; the scaling function and physical space representation are identical ; expansion coefficients are easily computed ; in certain respects they are more accurate ; the functions ( but not their derivatives ) can be computed without solving an eigenproblem . the price to be paid for these advantages is the loss of orthogonality , interpolating wavelets are only biorthogonal 本文主要的研究成果是把一維的某些結論推廣到高維,分為以下四個方面: ( 1 )使用二元拉格朗日插值法構造二元尺度函數和小波函數,使其具有緊支性、對稱性以及函數展開式的系數易于計算等優點。唯一的缺陷是缺乏正交性。
Based on the properties of bisymmetric matrices , a class of constrained inverse eigenproblem and associated approximation problem for bisymmetric matrices were essentially decomposed into the same kind of subproblems for real symmetric matrices with smaller dimensions , and the solutions of the two problems were obtained by applying the conclusions of real symmetric matrices 摘要根據雙對稱矩陣的性質,將雙對稱矩陣的一類約束逆特征值問題及其逼近問題分解成具有較小階數的實對稱矩陣的同類子問題,然后利用實對稱矩陣的結果導出雙對稱矩陣的這兩個問題的解。