6 APPENDIX: Technical results
6.1 A: The inverse function theorem
Lemma 6.1 (Contraction mapping principle) Let M be a complete metric space and
suppose T : M → M is a map such that
d(T x, T y) ≤ kd(x, y)
where k < 1. Then T has a unique fixed point.
Proof: Choose any point x0 , then
d(T m x0 , T n x0 ) ≤ k m d(x0 , T n−m x0 ) for n ≥ m
≤ k (d(x0 , T x0 ) + d(T x0 , T 2 x0 ) + . . . + d(T n−m−1 x0 , T n−m x0 ))
m
≤ k m (1 + k + . . . + k n−m−1 )d(x0 , T x0 )
km
≤ d(x0 , T x0 )
1−k
This is a Cauchy sequence, so completeness of M implies that it converges to x. Thus
x = lim T n x0 and so by continuity of T ,
T x = lim T n+1 x0 = x
For uniqueness, if T x = x and T y = y, then
d(x, y) = d(T x, T y) ≤ kd(x, y)
and so k < 1 implies d(x, y) = 0. 2
Theorem 6.2 (Inverse function theorem) Let U ⊆ Rn be an open set and f : U →
Rn a C ∞ function such that Dfa is invertible at a ∈ U . Then there exist neighbour-
hoods V, W of a and f (a) respectively such that f (V ) = W and f has a C ∞ inverse
on W .
Proof: By an affine transformation x 7→ Ax + b we can assume that a = 0 and
Dfa = I. Now consider g(x) = x − f (x). By construction Dg0 = 0 so by continuity
there exists r > 0 such that if kxk < 2r,
1
kDgx k <
2
92
It follows from the mean value theorem that
1
kg(x)k ≤ kxk
2
and so g maps the closed ball B̄(0, r) to B̄(0, r/2). Now consider
gy (x) = y + x − f (x)
(The choice of gy is made so that a fixed point gy (x) = x solves f (x) = y).
If now kyk ≤ r/2 and kxk ≤ r, then
1 1 1
kgy (x)k ≤ r + kg(x)k ≤ r + r = r
2 2 2
so gy maps the complete metric space M = B̄(0, r) to itself. Moreover
1
kgy (x1 ) − gy (x2 )k = kg(x1 ) − g(x2 )k ≤ kx1 − x2 k
2
if x1 , x2 ∈ B̄(0, r), and so gy is a contraction mapping. Applying Lemma 1 we have a
unique fixed point and hence an inverse ϕ = f −1 .
We need to show first that ϕ is continuous and secondly that it has derivatives of all
orders. From the definition of g and the mean value theorem,
kx1 − x2 k ≤ kf (x1 ) − f (x2 )k + kg(x1 ) − g(x2 )k
1
≤ kf (x1 ) − f (x2 )k + kx1 − x2 k
2
so
kx1 − x2 k ≤ 2kf (x1 ) − f (x2 )k
which is continuity for ϕ. It follows also from this inequality that if y1 = f (x1 ) and
y2 = f (x2 ) where y1 , y2 ∈ B(0, r/2) then x1 , x2 ∈ B̄(0, r), and so
kϕ(y1 ) − ϕ(y2 ) − (Dfx2 )−1 (y1 − y2 )k = kx1 − x2 − (Dfx2 )−1 (f (x1 ) − f (x2 ))k
≤ k(Dfx2 )−1 kkDfx2 (x1 − x2 ) − f (x1 ) + f (x2 )k
≤ Akx1 − x2 kR
where A is a bound on k(Dfx2 )−1 k and the function kx1 − x2 kR is the remainder
term in the definition of differentiability of f . But kx1 − x2 k ≤ 2ky1 − y2 k so as
y1 → y2 , x1 → x2 and hence R → 0, so ϕ is differentiable and moreover its derivative
is (Df )−1 .
Now we know the derivative of ϕ:
Dϕ = (Df )−1
so we see that it is continuous and has as many derivatives as f itself, so ϕ is C ∞ . 2
93
6.2 B: Existence of solutions of ordinary differential equa-
tions
Lemma 6.3 Let M be a complete metric space and T : M → M a map. If T n is a
contraction mapping, then T has a unique fixed point.
Proof: By the contraction mapping principle, T n has a unique fixed point x. We
also have
T n (T x) = T n+1 x = T (T n x) = T x
so T x is also a fixed point of T n . By uniqueness T x = x. 2
Theorem 6.4 Let f (t, x) be a continuous function on |t − t0 | ≤ a, kx − x0 k ≤ b and
suppose f satisfies a Lipschitz condition
kf (t, x1 ) − f (t, x2 )k ≤ kx1 − x2 k.
If M = sup |f (t, x)| and h = min(a, b/M ), then the differential equation
dx
= f (t, x), x(t0 ) = x0
dt
has a unique solution for |t − t0 | ≤ h.
Proof: Let Z t
(T x)(t) = x0 + f (s, x(s))ds
t0
Then T x is differentiable since f and x are continuous and if T x = x, x satisfies the
differential equation (differentiate the definition). We use the metric space
X = {x ∈ C([t0 − h, t0 + h], Rn ) : kx(t) − x0 k ≤ M h}
with the uniform metric
d(x1 , x2 ) = sup kx1 (t) − x2 (t)k
|t−t0 |≤h
which makes it complete. If x ∈ M , then T x ∈ M and we claim
ck
|T k x1 (t) − T k x2 (t)k ≤ |t − t0 |k d(x1 , x2 )
k!
94
For k = 0 this is clear, and in general we use induction to establish:
Z t
k k
kT x1 (t) − T x2 (t)k ≤ kf (s, T k−1 x1 (s) − f (s, T k−1 x2 (s)kds
t0
Z t
≤ c kT k−1 x1 (s) − T k−1 x2 (s)kds
t0
Z t
k
≤ (c /(k − 1)!) |s − t0 |k−1 ds d(x1 , x2 )
t0
k k
≤ (c /k!)|t − t0 | d(x1 , x2 )
So T n is a contraction mapping for large enough N , and the result follows. 2
Theorem 6.5 The solution above depends continuously on the initial data x0 .
Proof: Take h1 ≤ h and δ > 0 such that M h + δ ≤ b, and let
Y = {y ∈ C([t0 − h1 , t0 + h1 ] × B̄(x0 , δ); Rn : ky(t, x) − xk ≤ M h, y(t0 , x) = x}
which is a complete metric space as before. Now set
Z t
(T y)(t, x) = x + f (s, y(s, x))ds
t0
Since M h1 + δ ≤ b, T maps Y to Y and just as before T n is a contraction mapping
with a unique fixed point which satisfies
∂y
= f (t, y), y(t0 , x) = x
∂t
Since y is continuous in t and x this is what we need. 2
If f (t, x) is smooth then we need more work to prove that the solution to the equation
is smooth and smoothly dependent on parameters.
6.3 B’: Smooth dependence
Lemma 6.6 Let A(t, x), B(t, x) be continuous matrix-valued functions and take M ≥
supt,x kBk. The solutions of the linear differential equations
dξ(t, x)
= A(t, x)ξ(t, x), ξ(t0 , x) = a(x)
dt
dη(t, x)
= B(t, x)η(t, x), η(t0 , x) = b(x)
dt
95
satisfy
eM |t−t0 | − 1
sup kξ(t, x) − η(t, x)k ≤ CkA − Bk + ka − bkeM |t−t0 |
x M
where C is a constant depending only on A and a.
Proof: By the existence theorem we know how to find solutions as limits of ξn , ηn
where
Z t
ξk = a + Aξk−1 ds
t0
Z t
ηk = b + Bηk−1 ds
t0
Let gk (t) = supx kξk (t, x) − ηk (t, x)k and C = supk,x,t kξk k. Then
Z t
gn (t) ≤ ka − bk + CkA − Bk|t − t0 | + M gn−1 (s)ds
t0
Now define fn by f0 (t) = ka − bk and then inductively by
Z t
fn (t) = ka − bk + CkA − Bk|t − t0 | + M fn−1 (s)ds
t0
Comparing these two we see that fn ≥ gn . This is a contraction mapping, so that
fn → f with Z t
f (t) = ka − bk + CkA − Bk|t − t0 | + M f (s)ds
t0
and solving the corresponding differential equation we get
eM |t−t0 | − 1
f (t) = ka − bkeM |t−t0 | + CkA − Bk
M
As gn (t) ≤ fn (t),
sup kξn (t, x) − ηn (t, x)k ≤ fn (t)
x
and the theorem follows by letting n → ∞. 2
Theorem 6.7 If f is C k and
d
α(t, x) = f (t, α(t, x)), α(0, x) = x
dt
then α is also C k .
96
Proof: The hardest bit is k = 1. Assume f is C 1 so that ∂f /∂t and ∂f /∂xi exist
and are continuous. We must show that α is C 1 in all variables. If that were true,
then the matrix valued function λ where (λi = ∂α/∂xi ) would be the solution of the
differential equation
dλ
= Dx f (t, α)λ (14)
dt
so we shall solve this equation by the existence theorem and prove that the solution
is the derivative of α. Let F (s) = f (t, a + s(b − a)). Then
dF
= Dx f (t, a + s(b − a))(b − a)
ds
so Z 1
f (t, b) − f (t, a) = Dx f (t, a + s(b − a))(b − a)ds
0
But then
d
(α(t, x + y) − α(t, x)) = f (t, α(t, x + y)) − f (t, α(t, x))
dt Z 1
= Dx f (t, α(t, x) + s(α(t, x + y) − α(t, x)))(α(t, x + y) − α(t, x))ds
0
Let A(t, x) = Dx f (t, α(t, x)) and ξ(t, x) = λ(t, x)y and
Z 1
By (t, x) = Dx f (t, α(t, x)+s(α(t, x+y)−α(t, x)))ds, ηy (t, x) = α(t, x+y)−α(t, x)
0
Apply the previous lemma and we get
sup kλ(t, x)y − (α(t, x + y) − α(x))k = o(ky|)
|t|≤
and so Dx α = λ, which is continuous in (t, x). Since also dα/dt = f (t, α) this means
that α is C 1 in all variables.
To continue, suppose inductively that the theorem is true for k − 1, and f is C k .
Then A(t, x) = Dx f (t, α(t, x)) is C k−1 but since
dλ
= Aλ
dt
we have λ is C k−1 . Now Dx α = λ so the xi -derivatives of α are C k−1 . But also
dα/dt = f (t, α) is C k−1 too, so α is C k . 2
97
References
[1] P A Firby and C F Gardiner, “Surface topology”. Second edition. Ellis Horwood
Series: Mathematics and its Applications. Ellis Horwood, New York distributed
by Prentice Hall, Inc., Englewood Cliffs, NJ, 1991, ISBN 0-13-855321-1
[2] G K Francis and J R Weeks, Conway’s ZIP proof. Amer. Math. Monthly 106
(1999), 393–399.
[3] W S Massey, “A basic course in algebraic topology.” Graduate Texts in Mathe-
matics, 127. Springer-Verlag, New York, 1991. ISBN 0-387-97430-X
[4] H A Priestley, “Introduction to complex analysis”. Revised second edition. Oxford
University Press, Oxford, 2003. ISBN 0-19-852562-1
98