- Fréchet derivative
In

mathematics , the**Fréchet derivative**is aderivative defined onBanach space s. Named afterMaurice Fréchet , it is commonly used to formalize the concept of thefunctional derivative used widely inmathematical analysis , especiallyfunctional analysis . The Fréchet derivative also has applications tophysics , and in particular toquantum field theory . The Fréchet derivative should be contrasted to the more generalGâteaux derivative .Intuitively, the Fréchet derivative generalizes the idea of

linear approximation from functions of one variable to functions on Banach spaces.**Definition**Let "V" and "W" be Banach spaces, and $Usubset\; V$ be an

open subset of "V". A function "f" : "U" → "W" is called "Fréchet differentiable" at $x\; in\; U$ if there exists abounded linear operator $A\_x:V\; o\; W$ such that:$lim\_\{h\; o\; 0\}\; frac\{\; |\; f(x\; +\; h)\; -\; f(x)\; -\; A\_x(h)\; |\_\{W\}\; \}\{\; |h|\_\{V\}\; \}\; =\; 0.$

The limit here is meant in the usual sense of a

limit of a function defined on a metric space (see Functions on metric spaces), using "V" and "W" as the two metric spaces, and the above expression as the function of argument "h" in "V". As a consequence, it must exist for allsequence s $langle\; h\_n\; angle\_\{n=1\}^\{infty\}$ of non-zero elements of "V" which converge to the zero vector $h\_n\{\; ightarrow\}0.$ If the limit exists, we write $Df(x)=A\_x$ and call it the derivative of "f" at "x". A function "f" which is Fréchet differentiable for any point of "U", and whose derivative "Df"("x") is continuous in "x" on "U", is said to be C^{1}.This notion of derivative is a generalization of the ordinary derivative of a function on the

real number s "f" :**R**→**R**since the linear maps from**R**to**R**are just multiplication by a real number. In this case, "Df"("x") is the function $t\; mapsto\; tf\text{'}(x)$.**Properties**A function differentiable at a point is continuous at that point.

Differentiation is a linear operation in the following sense: if "f" and "g" are two maps "V" → "W" which are differentiable at "x", and "r" and "s" are scalars (two real or

complex number s), then "rf" + "sg" is differentiable at "x" with D("rf + sg")("x") = "r"D("f")("x") + "s"D("g")("x").The

chain rule is also valid in this context: if "f" : "U" → "Y" is differentiable at "x" in "U", and "g" : "Y" → "W" is differentiable at "y" = "f"("x"), then the composition "g" o "f" is differentiable in "x" and the derivative is the composition of the derivatives::$D(g\; circ\; f)(x)\; =\; D(g)(f(x))\; circ\; D(f)(x).$

**Finite dimensions**The Fréchet derivative in finite-dimensional spaces is the usual derivative.In particular, it is represented in coordinates by the Jacobi matrix.

Suppose that "f" is a map, "f":"U"⊂

**R**^{"n"}→**R**^{"m"}with "U" an open set. If "f" is Fréchet differentiable at a point "a" ∈ "U", then its derivative is:$Df(a)\; :\; mathbf\{R\}^n\; o\; mathbf\{R\}^m\; quadmbox\{with\}quad\; Df(a)(v)\; =\; J\_f(a)\; ,\; v,$where "J"_{"f"}("a") denotes the Jacobi matrix of "f" at "a".Furthermore, the partial derivatives of "f" are given by:$frac\{partial\; f\}\{partial\; x\_i\}(a)\; =\; Df(a)(e\_i)\; =\; J\_f(a)\; ,\; e\_i,$where {"e"

_{"i"}} is the canonical basis of**R**^{"n"}. Since the derivative is a linear function, we have for all vectors "h" ∈**R**^{"n"}that thedirectional derivative of "f" along "h" is given by:$Df(a)(h)\; =\; sum\_\{i=1\}^\{n\}\; h\_i\; frac\{partial\; f\}\{partial\; x\_i\}(a).$If all partial derivatives of "f" exist and are continuous, then "f" is Fréchet differentiable. The converse is not true: a function may be Fréchet differentiable and yet fail to have continuous partial derivatives.

**Relation to the Gâteaux derivative**A function "f" : "U" ⊂ "V" → "W" is called "Gâteaux differentiable" at $x\; in\; U$ if "f" has a directional derivative along all directions at "x". This means that there exists a function "g" : "V" → "W" such that

:$g(h)=lim\_\{t\; o\; 0\}\; frac\{\; f(x\; +\; th)\; -\; f(x)\; \}\{\; t\; \}$

for any chosen vector "h" in "V", and where "t" is from the scalar field associated with "V" (usually, "t" is real). If "f" is Fréchet differentiable at "x", it is also Gâteaux differentiable there, and "g" is just the linear operator "A" = "Df"("x"). However, not every Gâteaux differentiable function is Fréchet differentiable. If "f" is Gâteaux differentiable on an open set "U" ⊂ "X", then "f" is Fréchet differentiable if its Gâteaux derivative is linear and bounded at each point of "U" and the Gâteaux derivative is a continuous map "U" → "L"("X","Y").

For example, the real-valued function "f" of two real variables defined by :$f(x,\; y)=egin\{cases\}frac\{x^2y\}\{x^4+y^2\}\; mbox\{\; if\; \}\; (x,\; y)\; e\; (0,\; 0)\backslash 0\; mbox\{\; if\; \}\; (x,\; y)=(0,\; 0)end\{cases\}$is Gâteaux differentiable at (0, 0), with its derivative being:$g(a,\; b)=egin\{cases\}frac\{a^2\}\{b\}\; mbox\{\; if\; \}\; b\; e\; 0\; \backslash 0\; mbox\; \{\; if\; \}\; b=0end\{cases\}.$The function "g" is not a linear operator, so this function is not Fréchet differentiable.

In another situation, the function "f" given by :$f(x,\; y)=egin\{cases\}frac\{x^3y\}\{x^6+y^2\}\; mbox\{\; if\; \}\; (x,\; y)\; e\; (0,\; 0)\backslash 0\; mbox\{\; if\; \}\; (x,\; y)=(0,\; 0)end\{cases\}$is Gâteaux differentiable at (0, 0), with its derivative there being "g"("a", "b") = 0 for all ("a", "b"), which "is" a linear operator. However, "f" is not continuous at (0, 0) (one can see by approaching the origin along the curve ("t", "t"

^{3})) and therefore "f" cannot be Fréchet differentiable at the origin.The following example only works in infinite dimensions. Let "X" be a Banach space, and φ a

linear functional on "X" that is "discontinuous" at "x"=0. Let:$f(x)\; =\; |x|phi(x).,$

Then "f"("x") is Gâteaux differentiable at "x"=0 with derivative 0. However, "f"("x") is not Fréchet differentiable since the limit

:$lim\_\{x\; o\; 0\}phi(x)$

does not exist.

**Higher derivatives**If "f" is a differentiable function at all points in an open subset "U" of "V", it follows that its derivative :$D\; f\; :\; U\; o\; L(V,\; W)\; ,$

is a function from "U" to the space "L"("V", "W") of all bounded linear operators from "V" to "W". This function may as well have a derivative, the "second order derivative" of "f", which, by the definition of derivative, will be a map

:$D^2\; f\; :\; U\; o\; Lig(V,\; L(V,\; W)ig).$

To make it easier to work with second-order derivatives, the space on the right-hand side is identified with the Banach space "L"

^{2}("V"×"V", "W") of all continuousbilinear map s from "V" to "W". An element φ in "L"("V", "L"("V", "W")) is thus identified with ψ in "L"^{2}("V"×"V", "W") such that for all "x" and "y" in "V":$varphi(x)(y)=psi(x,\; y),$

(intuitively: a function φ linear in "x" with φ("x") linear in "y" is the same as a bilinear function ψ in "x" and "y").

One may differentiate

:$D^2\; f\; :\; U\; o\; L^2(V\; imes\; V,\; W)\; ,$

again, to obtain the "third order derivative", which at each point will be a "trilinear map", and so on. The "n"-th derivative will be a function

:$D^n\; f\; :\; U\; o\; L^n(V\; imes\; V\; imes\; cdots\; imes\; V,\; W),$

taking values in the Banach space of continuous

multilinear map s in "n" arguments from "V" to "W". Recursively, a function "f" is "n"+1 times differentiable on "U" if it is "n" times differentiable on "U" and for each "x" in "U" there exists a continuous multilinear map "A" of "n"+1 arguments such that the limit:$lim\_\{h\_\{n+1\}\; o\; 0\}\; frac\{\; |\; D^nf(x\; +\; h\_\{n+1\})(h\_1,\; h\_2,\; dots,\; h\_n)\; -\; D^nf(x)(h\_1,\; h\_2,\; dots,\; h\_n)\; -\; A(h\_1,\; h\_2,\; dots,\; h\_n,\; h\_\{n+1\})\; |\; \}\{\; |h\_\{n+1\}|\; \}\; =\; 0$

exists uniformly for "h"

_{1}, "h"_{2}, ..., "h"_{n}in bounded sets in "V". In that case, "A" is the "n"+1st derivative of "f" at "x".**See also***

Derivative (generalizations)

*Infinite-dimensional holomorphy **References*** Eric W Weisstein, Emma Previato, "Dictionary of Applied Math for Engineers and Scientists", CRC Press, 2002. ISBN 1584880538.

* James R. Munkres, "Analysis on Manifolds", ADDISON-WESLEY Publishing Company, 1990. ISBN 0-201-51035-9.

* H. Cartan, "Calcul Differentiel", Hermann, Paris, 1967.**External links*** B. A. Frigyik, S. Srivastava and M. R. Gupta, " [

*http://www.ee.washington.edu/research/guptalab/publications/functionalDerivativesIntroduction.pdf Introduction to Functional Derivatives*] ", UWEE Tech Report 2008-0001.

* http://www.probability.net. This webpage is mostly about basic probability and measure theory, but there is nice chapter about Frechet derivative in Banach spaces (chapter about Jacobian formula). All the results are given with proof.

*Wikimedia Foundation.
2010.*