Efficient algorithms for the inclusion of the inverse matrix using error- bounds for hyperpower methods (Q1179546): Difference between revisions

In this note the author continues his investigations concerning computable bounds for the inverse of a regular matrix \(A\). He generalizes previous results on error-bounds for hyperpower methods. The new bounds allow to establish a family of iterative methods for enclosing the inverse of \(A\). These methods employ interval operations. For practical purposes, i.e. in the presence of rounding errors, rounded interval operations have to be used (available for instance in PASCAL- SC). The efficiency index of some of the methods in the new family is higher than that of methods described by \textit{G. Alefeld} and the author [Introduction to interval computations (1983; Zbl 0552.65041)] and the author [Z. Angew. Math. Mech. 69, No. 3, 115-120 (1989; Zbl 0684.65027)]. An example, carried out in PASCAL-SC, illustrates the performance of the methods.