Riemannian Metrics on Smooth Manifolds
Introduction
The study of Riemannian geometry lies at the intersection of differential geometry, topology, and analysis. At its core, a Riemannian manifold provides a way to measure distances, angles, and curvature on smooth manifolds, extending our intuitive notions from Euclidean space to more complex geometric structures.
A smooth manifold is a topological space that locally resembles Euclidean space near each point, but may have a more complicated global structure. By equipping a manifold with a Riemannian metric, we gain the ability to define geometric concepts such as length, angles, geodesics, and curvature, which are essential tools in many areas of mathematics and physics.
The theory of Riemannian manifolds has profound applications ranging from general relativity to shape analysis in computer vision, and from mechanics to string theory. In this exposition, we will develop the fundamental concepts of Riemannian metrics and explore their properties and geometric implications.
Definitions
Definition 1: Riemannian Metric
A Riemannian metric on a smooth manifold M is a smooth assignment of an inner product g_p to the tangent space T_pM at each point p M. More formally, a Riemannian metric is a smooth symmetric 2-tensor field g on M such that for all p M, g_p is an inner product on T_pM.
Given a Riemannian metric g on a manifold M, we can define the length |X|_p of a tangent vector X T_pM by:
Similarly, the angle between two tangent vectors X, Y T_pM is given by:
In local coordinates (x^1, ..., x^n) on a neighborhood of p, the Riemannian metric can be expressed as:
where g_{ij}(x) = g(/x^i, /x^j) are smooth functions, and the matrix (g_{ij}(x)) is symmetric and positive definite for each x.
Definition 2: Riemannian Manifold
A Riemannian manifold is a pair (M, g), where M is a smooth manifold and g is a Riemannian metric on M.
Definition 3: Isometry
An isometry between Riemannian manifolds (M, g) and (N, h) is a diffeomorphism : M N such that for all p M and for all X, Y T_pM:
where d_p: T_pM T_{(p)}N is the differential of at p.
Examples
Example 1: Euclidean Space
The standard Euclidean metric on ^n is given by:
In terms of coordinates, this corresponds to the identity matrix (g_{ij}) where g_{ij} = _{ij} (the Kronecker delta). This metric gives us the familiar Euclidean distance and angle measures.
Example 2: Sphere
The n-sphere S^n embedded in ^{n+1} inherits a Riemannian metric from the ambient Euclidean space. In the usual spherical coordinates (_1, ..., _n) on S^n, this metric takes the form:
This metric on the sphere provides a measure of distance that corresponds to the length of paths along the surface of the sphere.
Example 3: Hyperbolic Space
There are several models for hyperbolic space H^n. In the Poincar ball model, where H^n is represented as the open unit ball {x ^n : |x| < 1}, the metric is given by:
This metric has constant negative curvature, unlike the Euclidean space (zero curvature) and the sphere (positive curvature).
Example 4: Surface of Revolution
Consider a surface obtained by rotating a curve (r(t), z(t)) around the z-axis in ^3. In coordinates (t, ), the metric takes the form:
This class of surfaces provides many concrete examples with interesting geometric properties.
Properties and Geometric Interpretations
Theorem 1: Existence of Riemannian Metrics
Every smooth manifold M admits a Riemannian metric. Moreover, there are infinitely many distinct Riemannian metrics on any given manifold of dimension at least 1.
The proof of this theorem relies on the existence of partitions of unity subordinate to a locally finite open cover of the manifold. Using this, one can construct Riemannian metrics locally (which is easy since each tangent space is a Euclidean space) and then patch them together smoothly.
Geodesics
Definition 4: Geodesic
A geodesic on a Riemannian manifold (M, g) is a curve whose velocity vector is parallel transported along itself, or equivalently, a curve that locally minimizes (or maximizes) length.
Geodesics are the generalization of straight lines in Euclidean space. In terms of local coordinates, they satisfy the geodesic equation:
where ^{k}_{ij} are the Christoffel symbols of the Levi-Civita connection associated with the metric g.
Curvature
One of the central concepts in Riemannian geometry is curvature, which measures how much a manifold deviates from being flat (isometric to Euclidean space). There are several ways to quantify curvature:
- Sectional curvature: Measures the curvature of two-dimensional subspaces of the tangent space, providing a complete description of the Riemannian curvature tensor.
- Ricci curvature: A trace of the sectional curvature, providing a less detailed but more computationally accessible measure, important in Einstein's field equations of general relativity.
- Scalar curvature: A further contraction of the Ricci curvature, giving a single number at each point that measures the average sectional curvature.
Theorem 2: Gauss-Bonnet Theorem
For a compact, orientable surface M with Riemannian metric g, the total Gaussian curvature satisfies:
where K is the Gaussian curvature, dA is the area element, and (M) is the Euler characteristic of M.
This beautiful theorem connects local geometric information (curvature) with global topological information (Euler characteristic), exemplifying the profound interplay between geometry and topology that characterizes Riemannian geometry.
Distance Function
Given a Riemannian manifold (M, g), we can define a distance function d: M M by letting d(p, q) be the infimum of the lengths of all piecewise smooth curves from p to q. The length of a smooth curve : [a, b] M is defined as:
This distance function turns M into a metric space, and if M is connected, this metric induces the same topology as the original manifold topology. This provides a crucial link between the smooth structure of M and its metric structure.
Applications
Riemannian geometry has far-reaching applications in various fields of mathematics and physics:
General Relativity
Einstein's theory of general relativity models spacetime as a 4-dimensional Lorentzian manifold, which is closely related to Riemannian manifolds (the only difference being that the metric has signature (3,1) instead of being positive definite). The curvature of this manifold is determined by the distribution of matter and energy through Einstein's field equations.
Shape Analysis and Computer Vision
In computer vision and shape analysis, Riemannian geometry provides tools for comparing and analyzing shapes. For instance, the space of shapes can be endowed with a Riemannian metric, allowing for meaningful measurement of distances between shapes and development of statistical methods for shape populations.
Optimization and Machine Learning
Riemannian optimization involves solving optimization problems where the variables are constrained to a Riemannian manifold. This has applications in machine learning, signal processing, and numerical linear algebra. For example, the space of covariance matrices forms a Riemannian manifold, and optimization algorithms can be developed to directly respect this structure.
Mechanics
In classical mechanics, the configuration space of a mechanical system often forms a manifold, and the kinetic energy defines a Riemannian metric on this space. The equations of motion can be seen in terms of geodesics with respect to this metric, providing a geometric formulation of mechanics.
String Theory
In string theory and other areas of theoretical physics, higher-dimensional manifolds with special geometric properties, such as Calabi-Yau manifolds with particular curvature conditions, play a crucial role in compactifying extra dimensions.
References
- D. B. O'Neill, Semi-Riemannian Geometry with Applications to Relativity, Academic Press, 1983.
- M. do Carmo, Riemannian Geometry, Birkhuser, 1992.
- J. M. Lee, Riemannian Manifolds: An Introduction to Curvature, Springer, 1997.
- S. Gallot, D. Hulin, and J. Lafontaine, Riemannian Geometry, Springer, 2004.
- W. Boothby, An Introduction to Differentiable Manifolds and Riemannian Geometry, Academic Press, 2003.
Second Order Smooth Variational Principle On Riemannian Manifolds and Reference File Downl...
