An acceleration of iterative processes for solving nonlinear equations (Q2572682)

An earlier method by the author [Mathematics and education in mathematics (Bulgarian) Sunny beach, 2003), 280--284, Union of Bulgarian Mathematicians, Sofia (2003)] for acceleration of algorithms for solving the nonlinear equation (1) \(f(x)=0\) having one and simple root is extended for more general cases. First, this approach is generalized for finding a multiple root \(\alpha\) of (1) with preliminary known multiplicity rate and after that it is proved for the case of simultaneous root-finding processes for multiple roots of algebraic polynomials. Some new formulas are as a result of the acceleration of already known formulae obtained, e.g., a modification of the Newton-secant formula for simultaneous approximation of all roots of an algebraic equation with arbitrary multiplicities is derived. Convergence properties of the presented methods are discussed and a few numerical experiments are performed in the MATLAB programming language.