基于共轭梯度法的思想，通过特殊的变形，建立了一类求矩阵方程AXA^T＋BYB^T=C的双对称最小二乘解的迭代算法．对任意的初始双对称矩阵．在没有舍人误差的情况下，经过有限步迭代得到它的双对称最小二乘解；在选取特殊的初始双对称矩阵时，能得到它的的极小范数双对称最小二乘解．另外，给定任意矩阵，利用此方法可得到它的最佳逼近双对称解，数值例子表明，这种方法是有效的．%On the base of conjugate gradient method, using special transformation, an iterative method is presented to solve the least squares bisymmetric solution pair of the linear matrix equation AXA^T+BYB^T=C.By this iterative method, the least squares bisymmetric solution pair can be obtained within finite iterative steps in the absence of roundoff errors, and minimum norm of the least squares solution pair can be obtained by choosing a special kind of initial matrix pair, In addition, the unique optimal approximation solution pair to the given matrices in Frobenius norm can be obtained. The given numerical examples demonstrate that the iterative methods are quite efficient.
展开▼
