3.5.1 Appendix: Real Analytic Functions
( \newcommand{\kernel}{\mathrm{null}\,}\)
Multi-index notation
The following multi-index notation simplifies many presentations of formulas. Let x=(x1,…,xn) and
u: Ω⊂R↦R1 (or Rm for systems).
The n-tuple of non-negative integers (including zero)
α=(α1,…,αn)
is called multi-index. Set
|α|=α1+…+αnα!=α1!α2!⋅…⋅αn!xα=xα11xα22⋅…⋅xαnn (for a monom)Dk=∂∂xkD=(D1,…,Dn)Du=(D1u,…,Dnu)≡∇u≡grad uDα=Dα11Dα22⋅…⋅Dαnn≡∂|α|∂xα11∂xα22…∂xαnn.
Define a partial order by
α≥β if and only if αi≥βi for all i.
Sometimes we use the notations
0=(0,0…,0), 1=(1,1…,1),
where
0, 1∈R.
Using this multi-index notion, we have
1.
(x+y)α=∑β,γβ+γ=αα!β!γ!xβyγ,
where x, y∈R and α, β, γ are multi-indices.
2. Taylor expansion for a polynomial f(x) of degree m:
f(x)=∑|α|≤m1α!(Dαf(0))xα,
here is Dαf(0):=(Dαf(x))|x=0.
3. Let x=(x1,…,xn) and m≥0 an integer, then
$$
(x_1+\ldots+x_n)^m=\sum_{|\alpha|=m}\frac{m!}{\alpha!}x^\alpha.
\]
4.
$$
\alpha!\le|\alpha|!\le n^{|\alpha|}\alpha!.
\]
5. Leibniz's rule:
$$
D^\alpha(fg)=
\sum_{β,γβ+γ=α}\frac{\alpha!}{\beta!\gamma!}(D^\beta f)(D^\gamma g).
\]
6.
Dβxα=α!(α−β)!xα−β if α≥β,Dβxα=0 otherwise.
7. Directional derivative:
dmdtmf(x+ty)=∑|α|=m|α|!α!(Dαf(x+ty))yα,
where x, y∈R and t∈R1.
8. Taylor's theorem: Let u∈Cm+1 in a neighborhood N(y) of y, then, if x∈N(y),
u(x)=∑|α|≤m1α!(Dαu(y))(x−y)α+Rm,
where
Rm=∑|α|=m+11α!(Dαu(y+δ(x−y)))xα, 0<δ<1,
δ=δ(u,m,x,y),
or
Rm=1m!∫10 (1−t)mΦ(m+1)(t) dt,
where Φ(t)=u(y+t(x−y)). It follows from 7. that
$$
R_m=(m+1)\sum_{|\alpha|=m+1}\frac{1}{\alpha!}\left(\int_0^1\ (1-t)D^\alpha u(y+t(x-y))\ dt\right)(x-y)^\alpha.
\]
9. Using multi-index notation, the general linear partial differential equation of order m can be written as
$$
\sum_{|\alpha|\le m}a_\alpha (x)D^\alpha u=f(x)\ \ \mbox{in}\ \Omega\subset\mathbb{R}.
\]
Power series
Here we collect some definitions and results for power series in R.
Definition. Let cα∈R1 (or ∈Rm). The series
∑αcα≡∞∑m=0(∑|α|=mcα)
is said to be convergent if
∑α|cα|≡∞∑m=0(∑|α|=m|cα|)
is convergent.
Remark. According to the above definition, a convergent series is absolutely convergent. It follows that we can rearrange the order of summation.
Using the above multi-index notation and keeping in mind that we can rearrange convergent series, we have
10. Let x∈R, then
∑αxα=n∏i=1(∞∑αi=0xαii)=1(1−x1)(1−x2)⋅…⋅(1−xn)=1(1−x)1,
provided |xi|<1 is satisfied for each i.
11. Assume x∈R and |x1|+|x2|+…+|xn|<1, then
∑α|α|!α!xα=∞∑j=0∑|α|=j|α|!α!xα=∞∑j=0(x1+…+xn)j=11−(x1+…+xn).
12. Let x∈R, |xi|<1 for all i, and β is a given multi-index. Then
∑α≥βα!(α−β)!xα−β=Dβ1(1−x)1=β!(1−x)1+β
13. Let x∈R and |x1|+…+|xn|<1. Then
∑α≥β|α|!(α−β)!xα−β=Dβ11−x1−…−xn=|β|!(1−x1−…−xn)1+|β| .
Consider the power series
∑αcαxα
and assume this series is convergent for a z∈R. Then, by definition,
μ:=∑α|cα||zα|<∞
and the series (3.34) is uniformly convergent for all x∈Q(z), where
$$
Q(z):\ \ |x_i|\le|z_i|\ \ \mbox{for all}\ \ i.
\]
Figure 3.5.1.1: Definition of D∈Q(z)
Thus the power series (3.34) defines a continuous function defined on Q(z), according to a theorem of Weierstrass.
The interior of Q(z) is not empty if and only if zi≠0 for all i, see Figure 3.5.1.1.
For given x in a fixed compact subset D of Q(z) there is a q, 0<q<1, such that
|xi|≤q|zi| for all i.
Set
$$
f(x)=\sum_\alpha c_\alpha x^\alpha.
\]
Proposition A1. (i) In every compact subset D of Q(z) one has f∈C∞(D) and
the formal differentiate series, that is ∑αDβcαxα, is uniformly convergent on the closure of D and is equal to Dβf.}
(ii)
|Dβf(x)|≤M|β|!r−|β| in D,
where
M=μ(1−q)n,r=(1−q)mini|zi|.
Proof. See F. John [10], p. 64. Or an exercise. Hint: Use formula 12. where x is replaced by (q,…,q).
Remark. From the proposition above it follows
$$
c_\alpha=\frac{1}{\alpha!}D^\alpha f(0).
\]
Definition. Assume f is defined on a domain Ω⊂R, then f is said to be {\it real analytic in y∈Ω} if there are cα∈R1 and if there is a neighborhood N(y) of y such that
f(x)=∑αcα(x−y)α
for all x∈N(y), and the series converges (absolutely) for each x∈N(y).
A function f is called {\it real analytic in Ω} if it is real analytic for each y∈Ω.
We will write f∈Cω(Ω) in the case that f is real analytic in the domain Ω.
A vector valued function f(x)=(f1(x),…,fm) is called real analytic if each coordinate is real analytic.
Proposition A2. (i) Let f∈Cω(Ω). Then f∈C∞(Ω).}
(ii)
Assume f∈Cω(Ω). Then for each y∈Ω there exists a neighborhood N(y) and positive constants M, r such that
f(x)=∑α1α!(Dαf(y))(x−y)α
for all x∈N(y), and the series converges (absolutely) for each x∈N(y), and
|Dβf(x)|≤M|β|!r−|β|.
The proof follows from Proposition A1.
An open set Ω∈R is called connected if Ω is not a union of two nonempty
open sets with empty intersection. An open set Ω∈R is connected if and only if its path connected, see [11], pp. 38, for example. We say that Ω is path connected if for any x,y∈Ω there is a continuous curve γ(t)∈Ω, 0≤t≤1, with γ(0)=x and γ(1)=y. From the theory of one complex variable we know that a continuation of an analytic function is uniquely determined. The same is true for real analytic functions.
Proposition A3. Assume f∈Cω(Ω) and Ω is connected. Then
f is uniquely determined if for one z∈Ω all Dαf(z) are known.
Proof. See F. John [10], p. 65. Suppose g,h∈Cω(Ω) and
Dαg(z)=Dαh(z) for every α. Set f=g−h and
Ω1={x∈Ω: Dαf(x)=0 for all α},Ω2={x∈Ω: Dαf(x)≠0 for at least one α}.
The set Ω2 is open since Dαf are continuous in Ω. The set Ω1 is also open since f(x)=0 in a neighbourhood of y∈Ω1. This follows from
f(x)=∑α1α!(Dαf(y))(x−y)α.
Since z∈Ω1, i. e., Ω1≠∅, it follows Ω2=∅.
◻
It was shown in Proposition A2 that derivatives of a real analytic function satisfy estimates.
On the other hand it follows, see the next proposition, that a function f∈C∞ is real analytic if these estimates are satisfied.
Definition. Let y∈Ω and M, r positive constants. Then f is said to be in the class CM,r(y) if f∈C∞ in a neighbourhood of y and if
|Dβf(y)|≤M|β|!r−|β|
for all β.
Proposition A4. f∈Cω(Ω) if and only if f∈C∞(Ω) and for every compact subset S⊂Ω there are positive constants M,r such that
f∈CM,r(y) for all y∈S.
Proof. See F. John [10], pp. 65-66. We will prove the local version of the proposition, that is, we show it for each fixed y∈Ω. The general version follows from Heine-Borel theorem. Because of Proposition A3 it remains to show that the Taylor series
∑α1α!Dαf(y)(x−y)α
converges (absolutely) in a neighborhood of y and that this series is equal to f(x).
Define a neighborhood of y by
Nd(y)={x∈Ω: |x1−y1|+…+|xn−yn|<d},
where d is a sufficiently small positive constant. Set Φ(t)=f(y+t(x−y)). The one-dimensional Taylor theorem says
f(x)=Φ(1)=j−1∑k=01k!Φ(k)(0)+rj,
where
rj=1(j−1)!∫10 (1−t)j−1Φ(j)(t) dt.
From formula 7. for directional derivatives it follows for x∈Nd(y) that
1j!djdtjΦ(t)=∑|α|=j1α!Dαf(y+t(x−y))(x−y)α.
From the assumption and the multinomial formula 3. we get for 0≤t≤1
|1j!djdtjΦ(t)|≤M∑|α|=j|α|!α!r−|α||(x−y)α|=Mr−j(|x1−y1|+…+|xn−yn|)j≤M(dr)j.
Choose d>0 such that d<r, then the Taylor series converges (absolutely) in Nd(y) and it is equal to f(x) since the remainder satisfies, see the above estimate,
$$
|r_j|=\left|\frac{1}{(j-1)!}\int_0^1\ (1-t)^{j-1}\Phi^j(t)\ dt\right|\le M\left(\frac{d}{r}\right)^j.
\]
◻
We recall that the notation f<<F (f is majorized by F) was defined in the previous section.
Proposition A5. (i) f=(f1,…,fm)∈CM,r(0) if and only if f<<(Φ,…,Φ), where
Φ(x)=Mrr−x1−…−xn .}
(ii) \(f\in C_{M,r}(0)\) and f(0)=0 if and only if
f<<(Φ−M,…,Φ−M),
where
Φ(x)=M(x1+…+xn)r−x1−…−xn .
Proof.
$$
D^\alpha\Phi(0)=M|\alpha|!r^{-|\alpha|}.
\]
◻
Remark. The definition of f<<F implies, trivially, that Dαf<<DαF.
The next proposition shows that compositions majorize if the involved functions majorize. More precisely, we have
Proposition A6. Let f, F: R↦Rm and g, G maps a neighborhood of 0∈Rm into Rp. Assume all functions f(x), F(x), g(u), G(u) are in C∞, f(0)=F(0)=0, f<<F and g<<G. Then
g(f(x))<<G(F(x)).}
Proof. See F. John [10], p. 68. Set
h(x)=g(f(x)), H(x)=G(F(x)).
For each coordinate hk of h we have, according to the chain rule,
Dαhk(0)=Pα(δβgl(0),Dγfj(0)),
where Pα are polynomials with non-negative integers as coefficients, Pα
are independent on g or f and δ:=(∂/∂u1,…,∂/∂um). Thus,
|Dαhk(0)|≤Pα(|δβgl(0)|,|Dγfj(0)|)≤Pα(δβGl(0),DγFj(0))=DαHk(0).
◻
Using this result and Proposition A4, which characterizes real analytic functions, it follows that compositions of real analytic functions are real analytic functions again.
Proposition A7. Assume f(x) and g(u) are real analytic, then g(f(x)) is real analytic if f(x) is in the domain of definition of g.
Proof. See F. John [10], p. 68. Assume that f maps a neighborhood of y∈R in Rm and g maps a neighborhood of v=f(y) in ${\mathbb R}^m$. Then f∈CM,r(y) and g∈Cμ,ρ(v) implies
h(x):=g(f(x))∈Cμ,ρr/(mM+ρ)(y).
Once one has shown this inclusion, the proposition follows from Proposition~A4. To show the inclusion, we set
h(y+x):=g(f(y+x))≡g(v+f(y+x)−f(x))=:g∗(f∗(x)),
where v=f(y) and
g∗(u):=g(v+u)∈Cμ,ρ(0)f∗(x):=f(y+x)−f(y)∈CM,r(0).
In the above formulas v, y are considered as fixed parameters. From Proposition~A5 it follows
f∗(x)<<(Φ−M,…,Φ−M)=:Fg∗(u)<<(Ψ,…,Ψ)=:G,
where
Φ(x)=Mrr−x1−x2−…−xnΨ(u)=μρρ−x1−x2−…−xn.
From Proposition A6 we get
h(y+x)<<(χ(x),…,χ(x))≡G(F),
where
χ(x)=μρρ−m(Φ(x)−M)=μρ(r−x1−…−xn)ρr−(ρ+mM)(x1+…+xn)<<μρrρr−(ρ+mM)(x1+…+xn)=μρr/(ρ+mM)ρr/(ρ+mM)−(x1+…xn).
See an exercise for the ''<<''-inequality.
◻
Contributors:
Integrated by Justin Marshall.