0% found this document useful (0 votes)
15 views3 pages

Understanding the Implicit Function Theorem

The Implicit Function Theorem states that under certain conditions involving Banach spaces and a continuously differentiable map, there exists a unique continuously differentiable function that satisfies specific properties related to the original map. If a continuous map also meets the conditions, it is equal to the function derived from the theorem. Additionally, the theorem ensures that if the original map is of class C^n or C^∞, the resulting function will maintain that class.

Uploaded by

Houssien Alarab
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
15 views3 pages

Understanding the Implicit Function Theorem

The Implicit Function Theorem states that under certain conditions involving Banach spaces and a continuously differentiable map, there exists a unique continuously differentiable function that satisfies specific properties related to the original map. If a continuous map also meets the conditions, it is equal to the function derived from the theorem. Additionally, the theorem ensures that if the original map is of class C^n or C^∞, the resulting function will maintain that class.

Uploaded by

Houssien Alarab
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd

The Implicit Function Theorem

Theorem : Let E, F, G be three Banach spaces, A ⊂ E × F an open and f : A 7−→ G a C 1


map. If (x0 , y0 ) ∈ A such that f (x0 , y0 ) = 0 and D2 f (x0 , y0 ) ∈ Isom(F, G), then there exists
an open U0 ⊂ E and a map u : U0 7−→ F with the properties

(1) x0 ∈ U0 and u(x0 ) = y0 ,


 
(2) (∀x ∈ U0 ) (x, u(x)) ∈ A and f (x, u(x)) = 0 ,
−1
(3) u is of class C 1 and Du(x) = − D2 f (x, u(x))

◦ D1 f (x, u(x)), for all x ∈ U0 .

Moreover, if U is an open connected set that satisfies x0 ∈ U ⊂ U0 and if v : U 7−→ F is a


continuous map and satisfies

(4) v(x0 ) = y0 ,
 
(5) (∀x ∈ U ) (x, v(x)) ∈ A and f (x, v(x)) = 0 ,

then v = u.

Remark : If f is of class C n then so is u.


If f is of class C ∞ then so is u.

Proof : Let T0 = D2 f (x0 , y0 ) and set g(x, y) = y−T0−1 f (x, y) and h(t, s) = g(x0 +t, y0 +s)−y0 .
The function h is defined near (0, 0) since A is open. Notice that g(x, y) = y ⇐⇒ f (x, y) = 0.
The map D2 f is continuous at (x0 , y0 ) and x 7−→ T0−1 f (x, y0 ) is continuous at x0 . Then there
exist α0 > 0 and β > 0 for which

1
kx − x0 k < α0 and ky − y0 k < β =⇒ kD2 f (x, y) − D2 f (x0 , y0 )k ≤
2kT0−1 k

β
and there exists γ > 0 for which kx − x0 k < γ =⇒ kT0−1 f (x, y0 )k < .
2

Set α = min(α0 , γ).


1 −1 β
For kx−x0 k < α and ky−y0 k < β we have kD2 f (x, y)−T0 k ≤ −1 and kT0 f (x, y0 )k < 2 .
2kT0 k
Let x be fixed and such that kx − x0 k < α, the function θx (y) = f (x, y) − T0 y is differentiable
near y0 and we have Dθx (y) = D2 f (x, y) − T0 .
1
If ky1 − y0 k < β and ky2 − y0 k < β then kf (x, y1 ) − f (x, y2 ) − T0 (y1 − y2 )k ≤ ky1 − y2 k.
2kT0−1 k

1
Then

1 1
ky1 − y2 − T0−1 f (x, y1 ) + T0−1 f (x, y2 )k ≤ ky1 − y2 k, i.e. kg(x, y1 ) − g(x, y2 )k ≤ ky1 − y2 k.
2 2
β
But g(x, y0 ) − y0 = −T0−1 f (x, y0 ) then kg(x, y0 ) − y0 k <
.
2
Then for every t ∈ B(0, α) and every s1 , s2 ∈ B(0, β) we have

1 1
kh(t, s1 )−h(t, s2 )k = kg(x0 +t, y0 +s1 )−y0 −g(x0 +t, y0 +s2 )+y0 k ≤ ky0 +s1 −y0 −s2 k = ks1 −s2 k
2 2
1
and kh(t, 0)k = kg(x0 + t, y0 ) − y0 k < β.
2
Then there exists a unique continuous map w : B(0, α) 7−→ B(0, β) for which w(t) = h(t, w(t)),
for all t ∈ B(0, α).
Consider u(x) = y0 + w(x − x0 ) ∈ B(y0 , β), for all x ∈ B(x0 , α). We have

h(x − x0 , w(x − x0 )) = w(x − x0 ) then g(x, y0 + w(x − x0 )) − y0 = u(x) − y0 ,

i.e. g(x, u(x)) = u(x) hence f (x, u(x)) = 0, for all x ∈ U0 = B(x0 , α).
On the other hand, we have u(x0 ) = y0 + w(0). Let us show that w(0) = 0.
Indeed, we have h(0, 0) = g(x0 , y0 ) − y0 = y0 − T0−1 f (x0 , y0 ) − y0 = −T0−1 0 = 0, then
1
kw(0)k = kh(0, w(0))k = kh(0, w(0)) − h(0, 0)k ≤ 2 kw(0)k, therefore w(0) = 0 and conse-
quently u(x0 ) = y0 .

In the sequel we can assume that D2 f (x, u(x)) ∈ Isom(F, G) for all x ∈ U0 . Indeed, consider
the mapping θ(x) = D2 f (x, u(x)),

θ : U0 7−→ U0 × B(y0 , β) ⊂ A 7−→ L(F, G)


x 7−→ (x, u(x))
(x, y) 7−→ D2 f (x, y)

So θ(x0 ) = T0 ∈ Isom(F, G). The mapping θ is continuous, then θ−1 (Isom(F, G)) 3 x0 is
open. Replace U0 by θ−1 (Isom(F, G)).

Let x, x + s ∈ U0 , so we have f (x, u(x)) = f (x + s, u(x + s)) = 0. Then ∀ε > 0, ∃r0 > 0
such that

ksk ≤ r0 and ktk ≤ r0 =⇒ kf (x+s, u(x)+t)−f (x, u(x))−D1 f (x, u(x))s−D2 f (x, u(x))tk ≤ ε(ktk+ksk).

2
Since u is continuous at x, for r0 > 0, there exists r00 > 0 such that

ksk ≤ r00 =⇒ ku(x + s) − u(x)k ≤ r0 .

For ksk ≤ r = min(r0 , r00 ) > 0, we have ku(x + s) − u(x)k ≤ r0 and ksk ≤ r0 .

Set t = u(x + s) − u(x), we have :

f (x + s, u(x + s)) − f (x, u(x)) − D1 f (x, u(x))s − D2 f (x, u(x))t


 −1 
= −D1 f (x, u(x))s − D2 f (x, u(x))t = −D2 f (x, u(x)) D2 f (x, u(x)) D1 f (x, u(x))s + t .

Then
 −1 −1
D2 f (x, u(x)) D1 f (x, u(x))s + t ≤ D2 f (x, u(x)) ε(ksk + ktk).
 −1
Call a = 2 D2 f (x, u(x)) ◦ D1 f (x, u(x)) + 1 > 0.
1  −1 1
If 0 < ε ≤  −1 we get D2 f (x, u(x)) D1 f (x, u(x))s + t ≤ 2 (ksk + ktk)
2 D2 f (x, u(x))
 −1 1
and ktk − D2 f (x, u(x)) D1 f (x, u(x))s ≤ (ksk + ktk). Then
2
 −1 1
ktk − D2 f (x, u(x)) D1 f (x, u(x))s ≤ (ksk + ktk).
2
 −1 a−1
Since D2 f (x, u(x)) D1 f (x, u(x))s ≤ ksk then
2
a−1  −1
− ksk ≤ − D2 f (x, u(x)) D1 f (x, u(x))s .
2
a−1 1
Therefore ktk − ksk ≤ (ksk + ktk) =⇒ ktk ≤ aksk.
2 2
Finally,
 −1  −1
t + D2 f (x, u(x)) D1 f (x, u(x))s ≤ εksk(1 + a) D2 f (x, u(x)) , then
 −1
Du(x) = − D2 f (x, u(x)) ◦ D1 f (x, u(x)) ∈ L(E, F ).

You might also like