This paper explores the complexity associated with solving the inverse Sturm-Liouville problem with Robin boundary conditions: given a sequence of eigenvalues and a sequence of norming constants, how many limits does a universal algorithm require to return the potential and boundary conditions? It is shown that if all but finitely many of the eigenvalues and norming constants coincide with those for the zero potential then the number of limits is zero, i.e. it is possible to retrieve the potential and boundary conditions precisely in finitely many steps. Otherwise, it is shown that this problem requires a single limit; moreover, if one has a priori control over how much the eigenvalues and norming constants differ from those of the zero-potential problem, and one knows that the average of the potential is zero, then the computation can be performed with complete error control. This is done in the spirit of the Solvability Complexity Index. All algorithms are provided explicitly along with numerical examples.
翻译:本文探索了与Robin边界条件有关的倒向Sturm-Liouville问题的复杂程度:根据一序列的egenvalue和一系列规范常数,一个通用算法需要多少限制才能返回潜力和边界条件? 该文件显示,如果除有限数量外,全部的egenvaly和规范常数都与零潜力的常数相吻合,那么限制的数量是零,也就是说,有可能在有限的许多步骤中准确检索潜力和边界条件。 否则,将显示这一问题需要一个单一的限制; 此外,如果一个人对eigenvality和规范常数与零潜能问题有多大程度的先验控制,而一个人知道潜力的平均值为零,那么计算就可以完全错误控制。这是本着可溶性复杂度指数的精神进行的。 所有算法都与数字示例一起明确提供。