Calculation of atomic energy level values
Calculation of atomic energy level values
复制标题
DOI：
10.1016/0010-4655(72)90048-3
复制
发表时间：
1984
影响因子：
6.3
通讯作者：
L. J. Radziemski;K. J. Fisher;D. W. Steinhaus;A. S. Goldman
中科院分区：
文献类型：
--
作者：
L. J. Radziemski;K. J. Fisher;D. W. Steinhaus;A. S. Goldman
已经编写了两种求解原子能级计算问题的最小二乘公式的方法。矩阵求逆方法能够处理具有多达 19,000 个分类的 285 x 1000 级数组。该方法的一个重要优点是计算了完整的方差-协方差矩阵，从而可以正确计算所计算的波数不确定性。迭代方法目前能够接受具有 20,000 个转换的 1000 x 1000 级别数组。它本质上能够计算更大数组的最小二乘答案，但缺点是无法轻松计算方差-协方差矩阵。高斯-赛德尔迭代法应用于级别计算问题已被证明是一个收敛迭代过程。
Two methods for solving the least-squares formulation of the atomicenergy-level calculation problem have been coded. The matrix-inversion method is capable of handling a 285 by 1000 level array with up to 19,000 classifications. An important advantage of this method is that the complete variance-covariance matrix is calculated, which leads to the correct computa tion of calculated wave-number uncertainties. The iterative method is presently capable of accepting a 1000 by 1000 level array with 20,000 transitions. It is inherently capable of computing the least-squares answers to even larger arrays, but has the disadvantage that the variance-covariance matrix cannot be easily calculated. The Gauss-Seidel iterative method as applied to the level calculation problem has been demonstrated to be a convergent iterative process.