6. Differential calculus of multivariate functions
Definition 1:
Let X and Y be normed spaces. A mapping f:E→Y of a set E⊂X into Y is differentiable at an interior point x∈E if there exists a continuous linear transformationL(x):X→Y such that
f(x+h)−f(x)=L(x)h+α(x;h)
where α(x;h)=o(h) as h→0,x+h∈E.
We often denote that L(x)as df(x),Df(x), or f′(x).
remark: If the condition for differentiability of the mapping f at some point a is written as
f(x)−f(a)=L(x)(x−a)+α(a;x)where α(a;x)=o(x−a) as x→a.
tangent space
We remark that strictly speaking f′(x)∈L(TXx;TYf(x)) here and h is a vector of the tangent space TXx. But parallel translation of a vector to any point x∈X is defined in a vector space, and this allows us to identify the tangent space TXx with the vector space X itself.
Consequently, after choosing a basis in X, we can extend it to all the tangent spaces TXx. This means that if, for example, X=Rm, Y=Rn,and the mappingf∈L(Rm;Rn) is given by the matrix (aij), then at every point x∈Rm the tangent mappingf′(x):TRxm→TRf(x)n will be given by the same matrix.
Chain Rule
If the mapping f:U→V is differentiable at a point x∈U⊂X, and the mapping g:V→Z is differentiable at f(x)=y∈V⊂Y, then the composition g∘f of these mappings is differentiable at x, and
(g∘f)′(x)=g′(f(x))∘f′(x)
Differentiation of the Inverse (#)
Let f:U→Y be a mapping that is continuous at x∈U⊂X and has an inverse f−1:V→X that is defined in a neighborhood of y=f(x) and continuous at that point. If the mapping f is differentiable at x and its tangent mapping f′(x)∈L(X;Y) has a continuous inverse [f′(x)]−1∈L(Y;X), then the mapping f−1 is differentiable at y=f(x) and
[f−1]′(f(x))=[f′(x)]−1
Partial derivative
Let U=U(a) be a neighborhoodof the point a∈X=X1×⋯×Xm in the direct product of the normed spaces X1,⋯,Xm , and let f:U→Y be a mapping of U into the normed space V. In this case
y=f(x)=f(x1,…,xm)
and hence, if we fix all the variables but xi in the above equation by settingxk=ak for k∈{1,…,m}\i, we obtain a function
f(a1,…,ai−1,xi,ai+1,…,am)=:φi(xi)
defined in some neighborhood Ui of ai in X.
We usually denote this partial derivative by one of the symbols
∂if(a),Dif(a),∂xi∂f(a),fxi′(a)
In accordance with these definitions Dif(a)∈L(Xi;Y). More precisely, Dif(a)∈L(TXi(ai);TY(f(a))).
Denote that h=(h1,…,hm)∈TX1(a1)×⋯×TXm(am)=TX(a), we have the equation
df(a)h=∂1f(a)h1+⋯+∂mf(a)hm
partial derivative of higher order
Consider U=U(x) be a neighborhoodof the point x=(x1,⋯,xn), Then if fhas partial derivative with respect to xi, then ∂xi∂fbecomes a new function which is defined in U, if:
∂xi1∂(∂xi2∂⋯(∂xik∂f)⋯)(x)
exist, we call it higher-order partial derivative, often denoted as:
∂xi1⋯∂xik∂kf(x)
Taylor's theorem
If a mapping f:U→Y from a neighborhood U=U(x) of a point x in a normed space X into a normed space Y has derivatives up to order n−1 inclusive in U and has an nth order derivative f(n)(x) at the point x, then
f(x+h)=f(x)+f′(x)h+⋯+n!1f(n)(x)hn+R(x;h)
R(x;h)=o(∣h∣n)when h→0.
We can expand the f(n)(x), such as Jacobian matrix: