Newton Method on Riemannian Manifolds: Covariant Alpha-Theory
نویسندگان
چکیده
Nf (z) = expz(−Df(z)f(z)) where expz : TzMn → Mn denotes the exponential map. When, instead of a mapping we consider a vector field X, in order to define Newton method, we resort to an object studied in differential geometry; namely, the covariant derivative of vector fields denoted here by DX. We let NX(z) = expz(−DX(z)X(z)). These definitions coincide with the usual one when Mn = Rn because expz(u) = z+ u and also because, in this context, the covariant derivative is just the usual derivative. The first to consider Newton method on a manifold is Rayleigh 1899 [7] who defined what we call today ”Rayleigh Quotient Iteration” which is in fact a Newton iteration for a vector field on the sphere. Then Shub 1986 [8] defined Newton’s method for the problem of finding the zeros of a vector field on a manifold and used retractions to send a neighborhood of the origin in the tangent space onto the manifold itself. In our paper we do not use general retractions but exponential maps. Udriste 1994 [19] studied Newton’s method to find the zeros of a gradient vector field defined on a Riemannian manifold; Owren and Welfert 1996 [6] defined Newton iteration for solving the equation F (x) = 0 where F is a map from a Lie group to its corresponding Lie algebra; Smith 1994 [18] and Edelman-Arias-Smith 1998 [3] developed Newton and conjugate gradient algorithms on the Grassmann and Stiefel manifolds. These authors define Newton’s method via the exponential map like we do here. Shub 1993 [9], Shub and Smale 1993-1996 [10], [11], [12], [13], [14], Malajovich 1994 [5], Dedieu and Shub 2000 [2] introduce and study Newton’s method on projective spaces and their products. Another paper about this subject is Adler-Dedieu-Margulies-Martens-Shub 2001 [1] where qualitative aspects of Newton method on Riemannian manifolds are investigated for both mappings and vector fields with an application to a geometric model for the human spine represented as a 18−tuple of 3×3 orthogonal matrices. Recently Ferreira-Svaiter [4] gave a Kantorovich like theorem for Newton method for vector fields defined on Riemannian manifolds.
منابع مشابه
Newton’s method on Riemannian manifolds: covariant alpha theory
In this paper, Smale’s α theory is generalized to the context of intrinsic Newton iteration on geodesically complete analytic Riemannian and Hermitian manifolds. Results are valid for analytic mappings from a manifold to a linear space of the same dimension, or for analytic vector fields on the manifold. The invariant γ is defined by means of high order covariant derivatives. Bounds on the size...
متن کاملRiemannian Convexity in Programming (II)
The book [7] emphasizes three relevant aspects: first, the fact that the notion of convexity is strongly metric-dependent either through geodesics or through the Riemannian connection; second that Riemannian convexity of functions is a coordinate-free concept, and consequently it can be easily connected with symbolic computation; third, that the Riemannian structure is involved essentially in f...
متن کاملAdaptive Numerical Treatment of Elliptic Systems on Manifolds
Adaptive multilevel finite element methods are developed and analyzed for certain elliptic systems arising in geometric analysis and general relativity. This class of nonlinear elliptic systems of tensor equations on manifolds is first reviewed, and then adaptive multilevel finite element methods for approximating solutions to this class of problems are considered in some detail. Two a posterio...
متن کاملACTION OF SEMISIMPLE ISOMERY GROUPS ON SOME RIEMANNIAN MANIFOLDS OF NONPOSITIVE CURVATURE
A manifold with a smooth action of a Lie group G is called G-manifold. In this paper we consider a complete Riemannian manifold M with the action of a closed and connected Lie subgroup G of the isometries. The dimension of the orbit space is called the cohomogeneity of the action. Manifolds having actions of cohomogeneity zero are called homogeneous. A classic theorem about Riemannian manifolds...
متن کاملExtended Newton’s Method for Mappings on Riemannian Manifolds with Values in a Cone
Robinson’s generalized Newton’s method for nonlinear functions with values in a cone is extended to mappings on Riemannian manifolds with values in a cone. When Df satisfies the L-average Lipschitz condition, we use the majorizing function technique to establish the semi-local quadratic convergence of the sequences generated by the extended Newton’s method. As applications, we also obtain Kanto...
متن کامل