Cauchy–Schwarz inequality

In mathematics, the Cauchy–Schwarz inequality, also known as the Cauchy–Bunyakovsky–Schwarz inequality, is a useful inequality in many mathematical fields, such as linear algebra, analysis, probability theory, vector algebra and other areas. It is considered to be one of the most important inequalities in all of mathematics.[1]

The inequality for sums was published by Augustin-Louis Cauchy (1821), while the corresponding inequality for integrals was first proved by Viktor Bunyakovsky (1859). The modern proof of the integral version was given by Hermann Schwarz (1888).[1]

Statement of the inequalityEdit

The Cauchy–Schwarz inequality states that for all vectors   and   of an inner product space it is true that

 

 

 

 

 

(Cauchy-Schwarz inequality [written using only the inner product])

where   is the inner product. Examples of inner products include the real and complex dot product; see the examples in inner product. Every inner product gives rise to a norm, called the canonical or induced norm, where the norm of a vector   is denoted and defined by:

 

so that this norm and the inner product are related by the defining condition   where   is always a non-negative real number (even if the inner product is complex-valued). By taking the square root of both sides of the above inequality, the Cauchy–Schwarz inequality can be written in its more familiar form:[2][3]

 

 

 

 

 

(Cauchy-Schwarz inequality [written using norm and inner product])

Moreover, the two sides are equal if and only if   and   are linearly dependent.[4][5]

Special casesEdit

Titu's lemma - Positive real numbersEdit

Titu's lemma (named after Titu Andreescu, also known as T2 lemma, Engel's form, or Sedrakyan's inequality) states that for positive reals, one has

 

It is a direct consequence of the Cauchy–Schwarz inequality, obtained upon substituting   and   This form is especially helpful when the inequality involves fractions where the numerator is a perfect square.

2 - The planeEdit

 
Cauchy-Schwarz inequality in a unit circle of the Euclidean plane

The real vector space   denotes the 2-dimensional plane. It is also the 2-dimensional Euclidean space where the inner product is the dot product. If   and   then the Cauchy–Schwarz inequality becomes:

 

where   is the angle between   and  

The form above is perhaps the easiest in which to understand the inequality, since the square of the cosine can be at most 1, which occurs when the vectors are in the same or opposite directions. It can also be restated in terms of the vector coordinates   and   as

 

where equality holds if and only if the vector   is in the same or opposite direction as the vector   or if one of them is the zero vector.

n - n-dimensional Euclidean spaceEdit

In Euclidean space   with the standard inner product, which is the dot product, the Cauchy–Schwarz inequality becomes:

 

The Cauchy–Schwarz inequality can be proved using only ideas from elementary algebra in this case. Consider the following quadratic polynomial in  

 

Since it is nonnegative, it has at most one real root for   hence its discriminant is less than or equal to zero. That is,

 

which yields the Cauchy–Schwarz inequality.

n - n-dimensional Complex spaceEdit

If   with   and   (where   and  ) and if the inner product on the vector space   is the canonical complex inner product (defined by  ), then the inequality may be restated more explicitly as follows (where the bar notation is used for complex conjugation):

 

That is,

 

L2Edit

For the inner product space of square-integrable complex-valued functions, the following inequality:

 

The Hölder inequality is a generalization of this.

ProofsEdit

There are many different proofs[6] of the Cauchy–Schwarz inequality other than those given below.[1][3] When consulting other sources, there are often two sources of confusion. First, some authors define ⟨⋅,⋅⟩ to be linear in the second argument rather than the first. Second, some proofs are only valid when the field is   and not  [7]

This section gives proofs of the following theorem:

Cauchy-Schwarz inequality — Let   and   be arbitrary vectors in an inner product space over the scalar field   where   is the field of real numbers   or complex numbers   Then

 

 

 

 

 

(Cauchy-Schwarz Inequality)

where in addition, equality holds in the Cauchy-Schwarz Inequality if and only if   and   are linearly dependent; explicitly, this means:

  if and only if one of   and   is a scalar multiple of the other.

 

 

 

 

(Characterization of Equality in Cauchy-Schwarz)

Moreover, if this equality holds and if   then  

In all of the proofs given below, the proof in the trivial case where at least one of the vectors is zero (or equivalently, in the case where  ) is the same. It is presented immediately below only once to reduce repetition. It also includes the easy part of the proof the above Equality Characterization; that is, it proves that if   and   are linearly dependent then  

Proof of the trivial parts: Case where a vector is   and also one direction of the Equality Characterization

By definition,   and   are linearly dependent if and only if one is a scalar multiple of the other. If   where   is some scalar then

 

which shows that equality holds in the Cauchy-Schwarz Inequality. The case where   for some scalar   is very similar, with the main difference between the complex conjugation of  :

 

If at least one of   and   is the zero vector then   and   are necessarily linearly dependent (just scalar multiply the non-zero vector by the number   to get the zero vector; e.g. if   then let   so that  ), which proves the converse of this characterization in this special case; that is, this shows that if at least one of   and   is   then the Equality Characterization holds.

If   which happens if and only if   then   and   so that in particular, the Cauchy-Schwarz inequality holds because both sides of it are   The proof in the case of   is identical.

Consequently, the Cauchy-Schwarz inequality only needs to be proven only for non-zero vectors and also only the non-trivial direction of the Equality Characterization must be shown.

Proof 1 —

The special case of   was proven above so it is henceforth assumed that   As is now shown, the Cauchy–Schwarz inequality (and the rest of the theorem) is an almost immediate corrollary of the following equality:

 [8]

 

 

 

 

(Eq. 1)

which is readily verified by elementarily expanding   (via the definition of the norm) and then simplifying.

Observing that the left hand side (LHS) of Eq. 1 is non-negative (which makes this also true of the right hand side (RHS)) proves that   from which the Cauchy-Schwarz Inequality follows (by taking the square root of both sides). If   then the RHS (and thus also the LHS) of Eq. 1 is   which is only possible if  ;[note 1] thus   which shows that   and   are linearly dependent.[8] Since the (trivial) converse was proved above, the proof of the theorem is complete. ⯀


Details of  's elementary expansion are now given for the interested reader. Let   and   so that   and   Then

 

Note that this expansion does not require   to be non-zero; however,   must be non-zero in order to divide both sides by   and to deduce the Cauchy-Schwarz inequality from it. Swapping   and   gives rise to:

 

and thus

 
Proof 2 —

The special case of   was proven above so it is henceforth assumed that   Let

 

It follows from the linearity of the inner product in its first argument that:

 

Therefore,   is a vector orthogonal to the vector   (Indeed,   is the projection of   onto the plane orthogonal to  ) We can thus apply the Pythagorean theorem to

 

which gives

 

The Cauchy–Schwarz inequality follows by multiplying by   and then taking the square root. Moreover, if the relation   in the above expression is actually an equality, then   and hence  ; the definition of   then establishes a relation of linear dependence between   and   The converse was proved at the beginning of this section, so the proof is complete. ⯀

Proof 3 —

The special case of   was proven above so it is henceforth assumed that   Let   be defined by

 

Then

 

Therefore,   or  

If the inequality holds as an equality, then   and so   thus   and   are linearly dependent. The converse was proved at the beginning of this section, so the proof is complete. ⯀

ApplicationsEdit

AnalysisEdit

In any inner product space, the triangle inequality is a consequence of the Cauchy–Schwarz inequality, as is now shown:

 

Taking square roots gives the triangle inequality:

 

The Cauchy–Schwarz inequality is used to prove that the inner product is a continuous function with respect to the topology induced by the inner product itself.[9][10]

GeometryEdit

The Cauchy–Schwarz inequality allows one to extend the notion of "angle between two vectors" to any real inner-product space by defining:[11][12]

 

The Cauchy–Schwarz inequality proves that this definition is sensible, by showing that the right-hand side lies in the interval [−1, 1] and justifies the notion that (real) Hilbert spaces are simply generalizations of the Euclidean space. It can also be used to define an angle in complex inner-product spaces, by taking the absolute value or the real part of the right-hand side,[13][14] as is done when extracting a metric from quantum fidelity.

Probability theoryEdit

Let   and   be random variables, then the covariance inequality:[15][16] is given by

 

After defining an inner product on the set of random variables using the expectation of their product,

 

the Cauchy–Schwarz inequality becomes

 

To prove the covariance inequality using the Cauchy–Schwarz inequality, let   and   then

 

where   denotes variance and   denotes covariance.

GeneralizationsEdit

Various generalizations of the Cauchy–Schwarz inequality exist. Hölder's inequality generalizes it to   norms. More generally, it can be interpreted as a special case of the definition of the norm of a linear operator on a Banach space (Namely, when the space is a Hilbert space). Further generalizations are in the context of operator theory, e.g. for operator-convex functions and operator algebras, where the domain and/or range are replaced by a C*-algebra or W*-algebra.

An inner product can be used to define a positive linear functional. For example, given a Hilbert space   being a finite measure, the standard inner product gives rise to a positive functional   by   Conversely, every positive linear functional   on   can be used to define an inner product   where   is the pointwise complex conjugate of   In this language, the Cauchy–Schwarz inequality becomes[17]

 

which extends verbatim to positive functionals on C*-algebras:

Cauchy–Schwarz inequality for positive functionals on C*-algebras[18][19] — If   is a positive linear functional on a C*-algebra   then for all    

The next two theorems are further examples in operator algebra.

Kadison–Schwarz inequality[20][21] (Named after Richard Kadison) — If   is a unital positive map, then for every normal element   in its domain, we have   and  

This extends the fact   when   is a linear functional. The case when   is self-adjoint, i.e.   is sometimes known as Kadison's inequality.

Cauchy-Schwarz inequality (Modified Schwarz inequality for 2-positive maps[22]) — For a 2-positive map   between C*-algebras, for all   in its domain,

 
 

Another generalization is a refinement obtained by interpolating between both sides the Cauchy-Schwarz inequality:

Callebaut's Inequality[23] — For reals  

 

This theorem can be deduced from Hölder's inequality.[24] There are also non commutative versions for operators and tensor products of matrices.[25]

See alsoEdit

NotesEdit

  1. ^ In fact, it follows immediately from Eq. 1 that   if and only if  

CitationsEdit

  1. ^ a b c Steele, J. Michael (2004). The Cauchy–Schwarz Master Class: an Introduction to the Art of Mathematical Inequalities. The Mathematical Association of America. p. 1. ISBN 978-0521546775. ...there is no doubt that this is one of the most widely used and most important inequalities in all of mathematics.
  2. ^ Strang, Gilbert (19 July 2005). "3.2". Linear Algebra and its Applications (4th ed.). Stamford, CT: Cengage Learning. pp. 154–155. ISBN 978-0030105678.
  3. ^ a b Hunter, John K.; Nachtergaele, Bruno (2001). Applied Analysis. World Scientific. ISBN 981-02-4191-7.
  4. ^ Bachmann, George; Narici, Lawrence; Beckenstein, Edward (2012-12-06). Fourier and Wavelet Analysis. Springer Science & Business Media. p. 14. ISBN 9781461205050.
  5. ^ Hassani, Sadri (1999). Mathematical Physics: A Modern Introduction to Its Foundations. Springer. p. 29. ISBN 0-387-98579-4. Equality holds iff <c|c>=0 or |c>=0. From the definition of |c>, we conclude that |a> and |b> must be proportional.
  6. ^ Wu, Hui-Hua; Wu, Shanhe (April 2009). "Various proofs of the Cauchy-Schwarz inequality" (PDF). Octogon Mathematical Magazine. 17 (1): 221–229. ISBN 978-973-88255-5-0. ISSN 1222-5657. Retrieved 18 May 2016.
  7. ^ Aliprantis, Charalambos D.; Border, Kim C. (2007-05-02). Infinite Dimensional Analysis: A Hitchhiker's Guide. Springer Science & Business Media. ISBN 9783540326960.
  8. ^ a b Halmos 1982, pp. 2, 167.
  9. ^ Bachman, George; Narici, Lawrence (2012-09-26). Functional Analysis. Courier Corporation. p. 141. ISBN 9780486136554.
  10. ^ Swartz, Charles (1994-02-21). Measure, Integration and Function Spaces. World Scientific. p. 236. ISBN 9789814502511.
  11. ^ Ricardo, Henry (2009-10-21). A Modern Introduction to Linear Algebra. CRC Press. p. 18. ISBN 9781439894613.
  12. ^ Banerjee, Sudipto; Roy, Anindya (2014-06-06). Linear Algebra and Matrix Analysis for Statistics. CRC Press. p. 181. ISBN 9781482248241.
  13. ^ Valenza, Robert J. (2012-12-06). Linear Algebra: An Introduction to Abstract Mathematics. Springer Science & Business Media. p. 146. ISBN 9781461209010.
  14. ^ Constantin, Adrian (2016-05-21). Fourier Analysis with Applications. Cambridge University Press. p. 74. ISBN 9781107044104.
  15. ^ Mukhopadhyay, Nitis (2000-03-22). Probability and Statistical Inference. CRC Press. p. 150. ISBN 9780824703790.
  16. ^ Keener, Robert W. (2010-09-08). Theoretical Statistics: Topics for a Core Course. Springer Science & Business Media. p. 71. ISBN 9780387938394.
  17. ^ Faria, Edson de; Melo, Welington de (2010-08-12). Mathematical Aspects of Quantum Field Theory. Cambridge University Press. p. 273. ISBN 9781139489805.
  18. ^ Lin, Huaxin (2001-01-01). An Introduction to the Classification of Amenable C*-algebras. World Scientific. p. 27. ISBN 9789812799883.
  19. ^ Arveson, W. (2012-12-06). An Invitation to C*-Algebras. Springer Science & Business Media. p. 28. ISBN 9781461263715.
  20. ^ Størmer, Erling (2012-12-13). Positive Linear Maps of Operator Algebras. Springer Monographs in Mathematics. Springer Science & Business Media. ISBN 9783642343698.
  21. ^ Kadison, Richard V. (1952-01-01). "A Generalized Schwarz Inequality and Algebraic Invariants for Operator Algebras". Annals of Mathematics. 56 (3): 494–503. doi:10.2307/1969657. JSTOR 1969657.
  22. ^ Paulsen, Vern (2002). Completely Bounded Maps and Operator Algebras. Cambridge Studies in Advanced Mathematics. 78. Cambridge University Press. p. 40. ISBN 9780521816694.
  23. ^ Callebaut, D.K. (1965). "Generalization of the Cauchy–Schwarz inequality". J. Math. Anal. Appl. 12 (3): 491–494. doi:10.1016/0022-247X(65)90016-8.
  24. ^ Callebaut's inequality. Entry in the AoPS Wiki.
  25. ^ Moslehian, M.S.; Matharu, J.S.; Aujla, J.S. (2011). "Non-commutative Callebaut inequality". arXiv:1112.3003 [math.FA].

ReferencesEdit

External linksEdit