Applying twice an idea of Hernández and Rubio (2002) for constructing a
one-parameter family of secant-like methods, we define a two-parameter family of
secant-like methods for solving nonlinear systems of equations. We analyze the
efficiency of this new family and conclude that the Kurchatov method, which is one
member of the family, is the most efficient. We illustrate this with Troesch’s
problem.
Keywords
nonlinear equations, iterative methods, divided difference,
secant method, Kurchatov method, secant-like method, order
of convergence, efficiency index, computational efficiency