In mathematics, the Hilbert projection theorem is a famous result of convex analysis that says that for every vector in a Hilbert space and every nonempty closed convex there exists a unique vector for which is minimized over the vectors ; that is, such that for every
Finite dimensional case
Some intuition for the theorem can be obtained by considering the first order condition of the optimization problem.
Consider a finite dimensional real Hilbert space with a subspace and a point If is a minimizer or minimum point of the function defined by (which is the same as the minimum point of ), then derivative must be zero at
In matrix derivative notation
Since is a vector in that represents an arbitrary tangent direction, it follows that must be orthogonal to every vector in
Statement
Detailed elementary proof
Proof by reduction to a special case
It suffices to prove the theorem in the case of because the general case follows from the statement below by replacing with
Consequences
Properties
Expression as a global minimum
The statement and conclusion of the Hilbert projection theorem can be expressed in terms of global minimums of the followings functions. Their notation will also be used to simplify certain statements.
Given a non-empty subset and some define a function
A global minimum point of if one exists, is any point in such that
in which case is equal to the global minimum value of the function which is:
Effects of translations and scalings
When this global minimum point exists and is unique then denote it by explicitly, the defining properties of (if it exists) are:
The Hilbert projection theorem guarantees that this unique minimum point exists whenever is a non-empty closed and convex subset of a Hilbert space.
However, such a minimum point can also exist in non-convex or non-closed subsets as well; for instance, just as long is is non-empty, if then
If is a non-empty subset, is any scalar, and are any vectors then
which implies:
Examples
The following counter-example demonstrates a continuous linear isomorphism for which
Endow with the dot product, let and for every real let be the line of slope through the origin, where it is readily verified that
Pick a real number and define by (so this map scales the coordinate by while leaving the coordinate unchanged).
Then is an invertible continuous linear operator that satisfies and
so that and
Consequently, if with and if then
See also
Orthogonal complement – Concept in linear algebra
Orthogonal projection – Idempotent linear transformation from a vector space to itselfPages displaying short descriptions of redirect targets
Orthogonality principle – Condition for optimality of Bayesian estimator
Riesz representation theorem – Theorem about the dual of a Hilbert space
Notes
References
Bibliography
Rudin, Walter (1987). Real and Complex Analysis (Third ed.).
Rudin, Walter (1991). Functional Analysis. International Series in Pure and Applied Mathematics. Vol. 8 (Second ed.). New York, NY: McGraw-Hill Science/Engineering/Math. ISBN 978-0-07-054236-5. OCLC 21163277.