Higher-order iterative methods for approximating zeros of analytic functions (Q1186915)

This paper presents iterative methods, which have rapid convergence in both floating point and circular arithmetic, for simultaneously approximating simple zeros of the analytic function \(\Phi(z)\). Basic single-step and total-step methods are considered, together with improvements based on the Newton \((N(z))\) and Halley \((H(z))\) corrections, respectively given by \(N(z)=\Phi(z)/\Phi'(z)\), \(H(z)=[(\Phi'(z)/\Phi(z))-(\Phi''(z)/2\Phi'(z))]^{-1}\). For the single-step methods under consideration, a table shows lower bounds on the \(R\)-order of each of these methods as a function of the number of zeros and demonstrates that the \(R\)-order lies between 4 and 7. Hybrid algorithms, combining the efficiency of floating-point iterative methods with accuracy control provided by interval arithmetic, are also considered.