On the efficiency of the secant method and the Newton method (Q677751)

An abstract discussion on the efficiency (in terms of the amount of arithmetic operations) of iterative methods for solving \(f(x)=0\) for \(f:\mathbb{R}^k\to\mathbb{R}^k\) is given. Under appropriate assumptions of smoothness of \(f\), three methods are considered: the standard chord (secant) method, a variant of the secant method given previously by the author, and the standard Newton method. According to a previous result of the author, the second method has convergence of order 2 under the usual hypotheses of the Newton-Kantorovich theory. On the other hand, the second method is shown to be more efficient than the first under the efficiency criteria discussed in the paper.