Newton's method on Lie groups (Q1034964): Difference between revisions

The article deals with the Newton-Kantorovich method of approximate solving nonlinear equations on Lie groups. In this case standard (modified) Newton-Kantorovich approximations can be written in the form \[ \begin{aligned} x_{n+1} & = x_n \cdot \exp (df_{x_n}^{-1} \circ f(x_n)) \qquad (n \geq 0)\\ (x_{n+1} & = x_n \cdot \exp (df_{x_0}^{-1} \circ f(x_n)) \qquad (n \geq 0));\end{aligned} \] here \(\exp\) is the exponential map of the Lie group \(G\) and \(f: G \to T_eG\) is a given \(C^1\)-mapping (\(T_eG\) is the tangent space of \(G\) at \(e\)). The main results are obtained under the additional assumption that \(G\) is Abelian. Two results are basic: a local theorem about the quadratic convergence of Newton-Kantorovich approximations to the solution of the equation \(f(x) = 0\) under the center Lipschitz condition for \(df_{x_0}^{-1}df\) and the semilocal theorem about the convergence of Newton-Kantorovich approximations and corresponding error estimates under the Lipschitz condition for \(df_{x_0}^{-1}df\) in a suitable ball at the center \(x_0\). Both theorems and their proofs are modifications of the arguments that the author systematically uses in analysis of Newton-Kantorovich method of solving nonlinear equations in Banach spaces. In the end of the article the modified Newton-Kantorovich approximations are considered.