Riemannian convexity of functionals

This paper extends the Riemannian convexity concept to action functionals defined by multiple integrals associated to Lagrangian differential forms on first order jet bundles. The main results of this paper are based on the geodesic deformations theory and their impact on functionals in Riemannian setting. They include the basic properties of Riemannian convex functionals, the Riemannian convexity of functionals associated to differential m-forms or to Lagrangians of class C ¹ respectively C ², the generalization to invexity and geometric meaningful convex functionals. Riemannian convexity of functionals is the central ingredient for global optimization. We illustrate the novel features of this theory, as well as its versatility, by introducing new definitions, theorems and algorithms that bear upon the currently active subject of functionals in variational calculus and optimal control. In fact so deep rooted is the convexity notion that nonconvex problems are tackled by devising appropriate convex approximations.     

Displacement interpolations from a Hamiltonian point of view

One of the most well-known results in the theory of optimal transportation is the equivalence between the convexity of the entropy functional with respect to the Riemannian Wasserstein metric and the Ricci curvature lower bound of the underlying Riemannian manifold. There are also generalizations of this result to the Finsler manifolds and manifolds with a Ricci flow background. In this paper, we study displacement interpolations from the point of view of Hamiltonian systems and give a unifying approach to the above mentioned results.     

Monotone and Accretive Vector Fields on Riemannian Manifolds

The relationship between monotonicity and accretivity on Riemannian manifolds is studied in this paper and both concepts are proved to be equivalent in Hadamard manifolds. As a consequence an iterative method is obtained for approximating singularities of Lipschitz continuous, strongly monotone mappings. We also establish the equivalence between the strong convexity of functions and the strong monotonicity of its subdifferentials on Riemannian manifolds. These results are then applied to solve the minimization of convex functions on Riemannian manifolds.     

Proximal subgradient and a characterization of Lipschitz function on Riemannian manifolds

A characterization of Lipschitz behavior of functions defined on Riemannian manifolds is given in this paper. First, it is extended the concept of proximal subgradient and some results of proximal analysis from Hilbert space to Riemannian manifold setting. A technique introduced by Clarke, Stern and Wolenski [F.H. Clarke, R.J. Stern, P.R. Wolenski, Subgradient criteria for monotonicity, the Lipschitz condition, and convexity, Canad. J. Math. 45 (1993) 1167-1183], for generating proximal subgradients of functions defined on a Hilbert spaces, is also extended to Riemannian manifolds in order to provide that characterization. A number of examples of Lipschitz functions are presented so as to show that the Lipschitz behavior of functions defined on Riemannian manifolds depends on the Riemannian metric.     

Continuity of optimal transport maps and convexity of injectivity domains on small deformations of 𝕊2

Given a compact Riemannian manifold, we study the regularity of the optimal transport map between two probability measures with cost given by the squared Riemannian distance. Our strategy is to define a new form of the so‐called Ma‐Trudinger‐Wang condition and to show that this condition, together with the strict convexity on the nonfocal domains, implies the continuity of the optimal transport map. Moreover, our new condition, again combined with the strict convexity of the nonfocal domains, allows us to prove that all injectivity domains are strictly convex too. These results apply, for instance, on any small C⁴‐deformation of the 2‐sphere. © 2009 Wiley Periodicals, Inc.     

Sufficient criteria of convexity

In this article a survey of criteria of convexity of sets and hypersurfaces inR ⁿ, in Riemannian and infinite-dimensional linear spaces, is made. Included are the following sections: support property; local convexity; local support; curvature of the boundary; section by planes; uniqueness of closest points; systems of sets. Translated from Zapiski Nauchnykh Seminarov Leningradskogo Otdeleniya Matematicheskogo Instituta im. V. A. Steklova AN SSSR, Vol. 45, pp. 3–52, 1974.     

Isosceles triangles in Riemannian geometry – a characterization of the <Emphasis Type="Italic">n</Emphasis>-sphere

New characterization theorems for the linear Euclidean spaces and Euclidean spheres are given in terms of convexity and isosceles triangles in Riemannian manifolds.     

Strong displacement convexity on Riemannian manifolds

Ricci curvature bounds in Riemannian geometry are known to be equivalent to the weak convexity (convexity along at least one geodesic between any two points) of certain functionals in the space of probability measures. We prove that the weak convexity can be reinforced into strong (usual) convexity, thus solving a question left open in Lott and Villani (Ann of Math, to appear). C. Villani is member of the Institut Universitaire de France.     

Convex domains of Finsler and Riemannian manifolds

A detailed study of the notions of convexity for a hypersurface in a Finsler manifold is carried out. In particular, the infinitesimal and local notions of convexity are shown to be equivalent. Our approach differs from Bishop??s one in his classical result (Bishop, Indiana Univ Math J 24:169?C172, 1974) for the Riemannian case. Ours not only can be extended to the Finsler setting but it also reduces the typical requirements of differentiability for the metric and it yields consequences on the multiplicity of connecting geodesics in the convex domain defined by the hypersurface.     

Convexity of the Solution Set of a Pseudoconvex Inequality in Riemannian Manifolds

Some equivalent conditions for convexity of the solution set of a pseudoconvex inequality are presented. These conditions turn out to be very useful in characterizing the solution sets of optimization problems of pseudoconvex functions defined on Riemannian manifold.     

Isometric immersion into sphere

Local convexity and convexity of submanifolds in a general Riemannian manifold have been defined. The propblem of isometric immersion of a submanifold into a unit sphere as a locally convex submanifold has been considered. The result we have arrived at is also true for the case of convex immersion.     

Tangent cut loci on surfaces

Given a smooth compact Riemannian surface, we prove that if a suitable convexity assumption on the tangent focal cut loci is satisfied, then all injectivity domains are semiconvex.     

Riemannian proximal gradient methods

In the Euclidean setting the proximal gradient method and its accelerated variants are a class of efficient algorithms for optimization problems with decomposable objective. In this paper, we develop a Riemannian proximal gradient method (RPG) and its accelerated variant (ARPG) for similar problems but constrained on a manifold. The global convergence of RPG is established under mild assumptions, and the O(1/k) is also derived for RPG based on the notion of retraction convexity. If assuming the objective function obeys the Rimannian Kurdyka–?ojasiewicz (KL) property, it is further shown that the sequence generated by RPG converges to a single stationary point. As in the Euclidean setting, local convergence rate can be established if the objective function satisfies the Riemannian KL property with an exponent. Moreover, we show that the restriction of a semialgebraic function onto the Stiefel manifold satisfies the Riemannian KL property, which covers for example the well-known sparse PCA problem. Numerical experiments on random and synthetic data are conducted to test the performance of the proposed RPG and ARPG.

     

Stochastic systems in Riemannian manifolds

A formulation of stochastic systems in a Riemannian manifold is given by stochastic differential equations in the tangent bundle of the manifold. Brownian motion is constructed in a compact Riemannian manifold as well as the horizontal lift of this process to the bundle of orthonormal frames. The solution of some stochastic differential equations in the tangent bundle of the manifold is defined by the transformation of the measure for the manifold-valued Brownian motion by a suitable Radon-Nikodym derivative. Real-valued stochastic integrals are defined for this Brownian motion using parallelism along the Brownian paths. A stochastic control problem is formulated and solved for these stochastic systems where a suitable convexity condition is assumed.This research was supported by NSF Grants Nos. GK-32136, ENG-75-06562, and MCS-76-01695.The author wishes to thank D. Gromoll, J. Simons, and J. Thorpe for some helpful conversations on differential geometry.     

Lower estimates for the first eigenvalue of the Laplace operator on doubly connected domains in a Riemannian manifold

Explicit lower estimates for the first eigenvalue of the Laplace operator in doubly connected domains of a Riemannian manifold are obtained, without any assumption on the mean convexity of the boundary of the domain, assuming either an upper bound of the sectional curvature, a lower bound of the Ricci curvature, or in highly symmetric manifolds where the Laplacian of the distance function to a fixed point depends only on the distance. Asymptotic properties are also analyzed. In many cases our estimates improve the classical and more recent ones.     

On the second eigenvalue of the dirichlet laplacian

The multiplicity of the second eigenvalue of the Dirichlet Laplacian on smooth Riemannian surfaces with boundary that satisfy certain convexity condition is at most two. The proof is based on variational formulas for eigenvalues under the change of the domain.     

A convexity theorem for noncommutative gradient flows

In this survey we shall prove a convexity theorem for gradient actions of reductive Lie groups on Riemannian symmetric spaces. After studying general properties of gradient maps, this proof is established by (1) an explicit calculation on the hyperbolic plane followed by a transfer of the results to general reductive Lie groups, (2) a reduction to a problem on abelian spaces using Kostant's Convexity Theorem, (3) an application of Fenchel's Convexity Theorem. In the final section the theorem is applied to gradient actions on other homogeneous spaces and we show, that Hilgert's Convexity Theorem for moment maps can be derived from the results.     

Surfaces convexes fuchsiennes dans les espaces lorentziens à courbure constante

We show existence and uniqueness of the equivariant isometric immersions of Riemannian surfaces into Lorentz space-forms under conditions implying convexity, when we impose that the associated representations leave a point invariant.

Received: 2 December 1995     

Mass Transport and Variants of the Logarithmic Sobolev Inequality

We develop the optimal transportation approach to modified log-Sobolev inequalities and to isoperimetric inequalities. Various sufficient conditions for such inequalities are given. Some of them are new even in the classical log-Sobolev case. The idea behind many of these conditions is that measures with a non-convex potential may enjoy such functional inequalities provided they have a strong integrability property that balances the lack of convexity. In addition, several known criteria are recovered in a simple unified way by transportation methods and generalized to the Riemannian setting. The research of A.V. Kolesnikov was supported by RFBR 07-01-00536, DFG Grant 436 RUS 113/343/0 and GFEN 06-01-39003.     

The radial part of Brownian motion II. Its life and times on the cut locus

Summary This paper is a sequel to Kendall (1987), which explained how the Itô formula for the radial part of Brownian motionX on a Riemannian manifold can be extended to hold for all time including those times a whichX visits the cut locus. This extension consists of the subtraction of a correction term, a continuous predictable non-decreasing processL which changes only whenX visits the cut locus. In this paper we derive a representation onL in terms of measures of local time ofX on the cut locus. In analytic terms we compute an expression for the singular part of the Laplacian of the Riemannian distance function. The work uses a relationship of the Riemannian distance function to convexity, first described by Wu (1979) and applied to radial parts of -martingales in Kendall (1993).The first author's research was supported by a visiting fellowship awarded by the UK Science and Engineering Council, by travel funds provided by a European Community SCIENCE initiative, by the Max-Planck-Institute of Bonn, and by a grant from NSA