Laguerre-like inclusion method for polynomial zeros (Q1872969)

The author presents a new iterative method for the simultaneous computation of all zeroes of a polynomial that has only simple roots. The method has order four of convergence (we note that Newton's method has only order 2) and is based on complex circular interval arithmetic. The main advantage of the use of circular interval arithmetic is that rigorous error bounds are automatically available. The paper is well-written and contains both a careful error analysis and extensive numerical tests.