Norm (mathematics)

(Redirected from L² norm)

In mathematics, a norm is a function from a real or complex vector space to the non-negative real numbers that behaves in certain ways like the distance from the origin: it commutes with scaling, obeys a form of the triangle inequality, and is zero only at the origin. In particular, the Euclidean distance in a Euclidean space is defined by a norm on the associated Euclidean vector space, called the Euclidean norm, the 2-norm, or, sometimes, the magnitude or length of the vector. This norm can be defined as the square root of the inner product of a vector with itself.

A seminorm satisfies the first two properties of a norm, but may be zero for vectors other than the origin.[1] A vector space with a specified norm is called a normed vector space. In a similar manner, a vector space with a seminorm is called a seminormed vector space.

The term pseudonorm has been used for several related meanings. It may be a synonym of "seminorm".[1] A pseudonorm may satisfy the same axioms as a norm, with the equality replaced by an inequality "" in the homogeneity axiom.[2][dubiousdiscuss] It can also refer to a norm that can take infinite values,[3] or to certain functions parametrised by a directed set.[4]

Definition

edit

Given a vector space   over a subfield   of the complex numbers   a norm on   is a real-valued function   with the following properties, where   denotes the usual absolute value of a scalar  :[5]

  1. Subadditivity/Triangle inequality:   for all  
  2. Absolute homogeneity:   for all   and all scalars  
  3. Positive definiteness/positiveness[6]/Point-separating: for all   if   then  
    • Because property (2.) implies   some authors replace property (3.) with the equivalent condition: for every     if and only if  

A seminorm on   is a function   that has properties (1.) and (2.)[7] so that in particular, every norm is also a seminorm (and thus also a sublinear functional). However, there exist seminorms that are not norms. Properties (1.) and (2.) imply that if   is a norm (or more generally, a seminorm) then   and that   also has the following property:

  1. Non-negativity:[6]   for all  

Some authors include non-negativity as part of the definition of "norm", although this is not necessary. Although this article defined "positive" to be a synonym of "positive definite", some authors instead define "positive" to be a synonym of "non-negative";[8] these definitions are not equivalent.

Equivalent norms

edit

Suppose that   and   are two norms (or seminorms) on a vector space   Then   and   are called equivalent, if there exist two positive real constants   and   with   such that for every vector     The relation "  is equivalent to  " is reflexive, symmetric (  implies  ), and transitive and thus defines an equivalence relation on the set of all norms on   The norms   and   are equivalent if and only if they induce the same topology on  [9] Any two norms on a finite-dimensional space are equivalent but this does not extend to infinite-dimensional spaces.[9]

Notation

edit

If a norm   is given on a vector space   then the norm of a vector   is usually denoted by enclosing it within double vertical lines:   Such notation is also sometimes used if   is only a seminorm. For the length of a vector in Euclidean space (which is an example of a norm, as explained below), the notation   with single vertical lines is also widespread.

Examples

edit

Every (real or complex) vector space admits a norm: If   is a Hamel basis for a vector space   then the real-valued map that sends   (where all but finitely many of the scalars   are  ) to   is a norm on  [10] There are also a large number of norms that exhibit additional properties that make them useful for specific problems.

Absolute-value norm

edit

The absolute value   is a norm on the vector space formed by the real or complex numbers. The complex numbers form a one-dimensional vector space over themselves and a two-dimensional vector space over the reals; the absolute value is a norm for these two structures.

Any norm   on a one-dimensional vector space   is equivalent (up to scaling) to the absolute value norm, meaning that there is a norm-preserving isomorphism of vector spaces   where   is either   or   and norm-preserving means that   This isomorphism is given by sending   to a vector of norm   which exists since such a vector is obtained by multiplying any non-zero vector by the inverse of its norm.

Euclidean norm

edit

On the  -dimensional Euclidean space   the intuitive notion of length of the vector   is captured by the formula[11]  

This is the Euclidean norm, which gives the ordinary distance from the origin to the point X—a consequence of the Pythagorean theorem. This operation may also be referred to as "SRSS", which is an acronym for the square root of the sum of squares.[12]

The Euclidean norm is by far the most commonly used norm on  [11] but there are other norms on this vector space as will be shown below. However, all these norms are equivalent in the sense that they all define the same topology on finite-dimensional spaces.

The inner product of two vectors of a Euclidean vector space is the dot product of their coordinate vectors over an orthonormal basis. Hence, the Euclidean norm can be written in a coordinate-free way as  

The Euclidean norm is also called the quadratic norm,   norm,[13]   norm, 2-norm, or square norm; see   space. It defines a distance function called the Euclidean length,   distance, or   distance.

The set of vectors in   whose Euclidean norm is a given positive constant forms an  -sphere.

Euclidean norm of complex numbers

edit

The Euclidean norm of a complex number is the absolute value (also called the modulus) of it, if the complex plane is identified with the Euclidean plane   This identification of the complex number   as a vector in the Euclidean plane, makes the quantity   (as first suggested by Euler) the Euclidean norm associated with the complex number. For  , the norm can also be written as   where   is the complex conjugate of  

Quaternions and octonions

edit

There are exactly four Euclidean Hurwitz algebras over the real numbers. These are the real numbers   the complex numbers   the quaternions   and lastly the octonions   where the dimensions of these spaces over the real numbers are   respectively. The canonical norms on   and   are their absolute value functions, as discussed previously.

The canonical norm on   of quaternions is defined by   for every quaternion   in   This is the same as the Euclidean norm on   considered as the vector space   Similarly, the canonical norm on the octonions is just the Euclidean norm on  

Finite-dimensional complex normed spaces

edit

On an  -dimensional complex space   the most common norm is  

In this case, the norm can be expressed as the square root of the inner product of the vector and itself:   where   is represented as a column vector   and   denotes its conjugate transpose.

This formula is valid for any inner product space, including Euclidean and complex spaces. For complex spaces, the inner product is equivalent to the complex dot product. Hence the formula in this case can also be written using the following notation:  

Taxicab norm or Manhattan norm

edit

  The name relates to the distance a taxi has to drive in a rectangular street grid (like that of the New York borough of Manhattan) to get from the origin to the point  

The set of vectors whose 1-norm is a given constant forms the surface of a cross polytope, which has dimension equal to the dimension of the vector space minus 1. The Taxicab norm is also called the   norm. The distance derived from this norm is called the Manhattan distance or   distance.

The 1-norm is simply the sum of the absolute values of the columns.

In contrast,   is not a norm because it may yield negative results.

p-norm

edit

Let   be a real number. The  -norm (also called  -norm) of vector   is[11]   For   we get the taxicab norm, for   we get the Euclidean norm, and as   approaches   the  -norm approaches the infinity norm or maximum norm:   The  -norm is related to the generalized mean or power mean.

For   the  -norm is even induced by a canonical inner product   meaning that   for all vectors   This inner product can be expressed in terms of the norm by using the polarization identity. On   this inner product is the Euclidean inner product defined by   while for the space   associated with a measure space   which consists of all square-integrable functions, this inner product is  

This definition is still of some interest for   but the resulting function does not define a norm,[14] because it violates the triangle inequality. What is true for this case of   even in the measurable analog, is that the corresponding   class is a vector space, and it is also true that the function   (without  th root) defines a distance that makes   into a complete metric topological vector space. These spaces are of great interest in functional analysis, probability theory and harmonic analysis. However, aside from trivial cases, this topological vector space is not locally convex, and has no continuous non-zero linear forms. Thus the topological dual space contains only the zero functional.

The partial derivative of the  -norm is given by  

The derivative with respect to   therefore, is   where   denotes Hadamard product and   is used for absolute value of each component of the vector.

For the special case of   this becomes   or  

Maximum norm (special case of: infinity norm, uniform norm, or supremum norm)

edit
 
 

If   is some vector such that   then:  

The set of vectors whose infinity norm is a given constant,   forms the surface of a hypercube with edge length  

Zero norm

edit

In probability and functional analysis, the zero norm induces a complete metric topology for the space of measurable functions and for the F-space of sequences with F–norm  [15] Here we mean by F-norm some real-valued function   on an F-space with distance   such that   The F-norm described above is not a norm in the usual sense because it lacks the required homogeneity property.

Hamming distance of a vector from zero

edit

In metric geometry, the discrete metric takes the value one for distinct points and zero otherwise. When applied coordinate-wise to the elements of a vector space, the discrete distance defines the Hamming distance, which is important in coding and information theory. In the field of real or complex numbers, the distance of the discrete metric from zero is not homogeneous in the non-zero point; indeed, the distance from zero remains one as its non-zero argument approaches zero. However, the discrete distance of a number from zero does satisfy the other properties of a norm, namely the triangle inequality and positive definiteness. When applied component-wise to vectors, the discrete distance from zero behaves like a non-homogeneous "norm", which counts the number of non-zero components in its vector argument; again, this non-homogeneous "norm" is discontinuous.

In signal processing and statistics, David Donoho referred to the zero "norm" with quotation marks. Following Donoho's notation, the zero "norm" of   is simply the number of non-zero coordinates of   or the Hamming distance of the vector from zero. When this "norm" is localized to a bounded set, it is the limit of  -norms as   approaches 0. Of course, the zero "norm" is not truly a norm, because it is not positive homogeneous. Indeed, it is not even an F-norm in the sense described above, since it is discontinuous, jointly and severally, with respect to the scalar argument in scalar–vector multiplication and with respect to its vector argument. Abusing terminology, some engineers[who?] omit Donoho's quotation marks and inappropriately call the number-of-non-zeros function the   norm, echoing the notation for the Lebesgue space of measurable functions.

Infinite dimensions

edit

The generalization of the above norms to an infinite number of components leads to   and   spaces for   with norms

 

for complex-valued sequences and functions on   respectively, which can be further generalized (see Haar measure). These norms are also valid in the limit as  , giving a supremum norm, and are called   and  

Any inner product induces in a natural way the norm  

Other examples of infinite-dimensional normed vector spaces can be found in the Banach space article.

Generally, these norms do not give the same topologies. For example, an infinite-dimensional   space gives a strictly finer topology than an infinite-dimensional   space when  

Composite norms

edit

Other norms on   can be constructed by combining the above; for example   is a norm on  

For any norm and any injective linear transformation   we can define a new norm of   equal to   In 2D, with   a rotation by 45° and a suitable scaling, this changes the taxicab norm into the maximum norm. Each   applied to the taxicab norm, up to inversion and interchanging of axes, gives a different unit ball: a parallelogram of a particular shape, size, and orientation.

In 3D, this is similar but different for the 1-norm (octahedrons) and the maximum norm (prisms with parallelogram base).

There are examples of norms that are not defined by "entrywise" formulas. For instance, the Minkowski functional of a centrally-symmetric convex body in   (centered at zero) defines a norm on   (see § Classification of seminorms: absolutely convex absorbing sets below).

All the above formulas also yield norms on   without modification.

There are also norms on spaces of matrices (with real or complex entries), the so-called matrix norms.

In abstract algebra

edit

Let   be a finite extension of a field   of inseparable degree   and let   have algebraic closure   If the distinct embeddings of   are   then the Galois-theoretic norm of an element   is the value   As that function is homogeneous of degree  , the Galois-theoretic norm is not a norm in the sense of this article. However, the  -th root of the norm (assuming that concept makes sense) is a norm.[16]

Composition algebras

edit

The concept of norm   in composition algebras does not share the usual properties of a norm since null vectors are allowed. A composition algebra   consists of an algebra over a field   an involution   and a quadratic form   called the "norm".

The characteristic feature of composition algebras is the homomorphism property of  : for the product   of two elements   and   of the composition algebra, its norm satisfies   In the case of division algebras       and   the composition algebra norm is the square of the norm discussed above. In those cases the norm is a definite quadratic form. In the split algebras the norm is an isotropic quadratic form.

Properties

edit

For any norm   on a vector space   the reverse triangle inequality holds:   If   is a continuous linear map between normed spaces, then the norm of   and the norm of the transpose of   are equal.[17]

For the   norms, we have Hölder's inequality[18]   A special case of this is the Cauchy–Schwarz inequality:[18]  

 
Illustrations of unit circles in different norms.

Every norm is a seminorm and thus satisfies all properties of the latter. In turn, every seminorm is a sublinear function and thus satisfies all properties of the latter. In particular, every norm is a convex function.

Equivalence

edit

The concept of unit circle (the set of all vectors of norm 1) is different in different norms: for the 1-norm, the unit circle is a square oriented as a diamond; for the 2-norm (Euclidean norm), it is the well-known unit circle; while for the infinity norm, it is an axis-aligned square. For any  -norm, it is a superellipse with congruent axes (see the accompanying illustration). Due to the definition of the norm, the unit circle must be convex and centrally symmetric (therefore, for example, the unit ball may be a rectangle but cannot be a triangle, and   for a  -norm).

In terms of the vector space, the seminorm defines a topology on the space, and this is a Hausdorff topology precisely when the seminorm can distinguish between distinct vectors, which is again equivalent to the seminorm being a norm. The topology thus defined (by either a norm or a seminorm) can be understood either in terms of sequences or open sets. A sequence of vectors   is said to converge in norm to   if   as   Equivalently, the topology consists of all sets that can be represented as a union of open balls. If   is a normed space then[19]  

Two norms   and   on a vector space   are called equivalent if they induce the same topology,[9] which happens if and only if there exist positive real numbers   and   such that for all     For instance, if   on   then[20]  

In particular,       That is,   If the vector space is a finite-dimensional real or complex one, all norms are equivalent. On the other hand, in the case of infinite-dimensional vector spaces, not all norms are equivalent.

Equivalent norms define the same notions of continuity and convergence and for many purposes do not need to be distinguished. To be more precise the uniform structure defined by equivalent norms on the vector space is uniformly isomorphic.

Classification of seminorms: absolutely convex absorbing sets

edit

All seminorms on a vector space   can be classified in terms of absolutely convex absorbing subsets   of   To each such subset corresponds a seminorm   called the gauge of   defined as   where   is the infimum, with the property that   Conversely:

Any locally convex topological vector space has a local basis consisting of absolutely convex sets. A common method to construct such a basis is to use a family   of seminorms   that separates points: the collection of all finite intersections of sets   turns the space into a locally convex topological vector space so that every p is continuous.

Such a method is used to design weak and weak* topologies.

norm case:

Suppose now that   contains a single   since   is separating,   is a norm, and   is its open unit ball. Then   is an absolutely convex bounded neighbourhood of 0, and   is continuous.
The converse is due to Andrey Kolmogorov: any locally convex and locally bounded topological vector space is normable. Precisely:
If   is an absolutely convex bounded neighbourhood of 0, the gauge   (so that   is a norm.

See also

edit

References

edit
  1. ^ a b Knapp, A.W. (2005). Basic Real Analysis. Birkhäuser. p. [1]. ISBN 978-0-817-63250-2.
  2. ^ "Pseudo-norm - Encyclopedia of Mathematics". encyclopediaofmath.org. Retrieved 2022-05-12.
  3. ^ "Pseudonorm". www.spektrum.de (in German). Retrieved 2022-05-12.
  4. ^ Hyers, D. H. (1939-09-01). "Pseudo-normed linear spaces and Abelian groups". Duke Mathematical Journal. 5 (3). doi:10.1215/s0012-7094-39-00551-x. ISSN 0012-7094.
  5. ^ Pugh, C.C. (2015). Real Mathematical Analysis. Springer. p. page 28. ISBN 978-3-319-17770-0. Prugovečki, E. (1981). Quantum Mechanics in Hilbert Space. p. page 20.
  6. ^ a b Kubrusly 2011, p. 200.
  7. ^ Rudin, W. (1991). Functional Analysis. p. 25.
  8. ^ Narici & Beckenstein 2011, pp. 120–121.
  9. ^ a b c Conrad, Keith. "Equivalence of norms" (PDF). kconrad.math.uconn.edu. Retrieved September 7, 2020.
  10. ^ Wilansky 2013, pp. 20–21.
  11. ^ a b c Weisstein, Eric W. "Vector Norm". mathworld.wolfram.com. Retrieved 2020-08-24.
  12. ^ Chopra, Anil (2012). Dynamics of Structures, 4th Ed. Prentice-Hall. ISBN 978-0-13-285803-8.
  13. ^ Weisstein, Eric W. "Norm". mathworld.wolfram.com. Retrieved 2020-08-24.
  14. ^ Except in   where it coincides with the Euclidean norm, and   where it is trivial.
  15. ^ Rolewicz, Stefan (1987), Functional analysis and control theory: Linear systems, Mathematics and its Applications (East European Series), vol. 29 (Translated from the Polish by Ewa Bednarczuk ed.), Dordrecht; Warsaw: D. Reidel Publishing Co.; PWN—Polish Scientific Publishers, pp. xvi, 524, doi:10.1007/978-94-015-7758-8, ISBN 90-277-2186-6, MR 0920371, OCLC 13064804
  16. ^ Lang, Serge (2002) [1993]. Algebra (Revised 3rd ed.). New York: Springer Verlag. p. 284. ISBN 0-387-95385-X.
  17. ^ Trèves 2006, pp. 242–243.
  18. ^ a b Golub, Gene; Van Loan, Charles F. (1996). Matrix Computations (Third ed.). Baltimore: The Johns Hopkins University Press. p. 53. ISBN 0-8018-5413-X.
  19. ^ Narici & Beckenstein 2011, pp. 107–113.
  20. ^ "Relation between p-norms". Mathematics Stack Exchange.

Bibliography

edit