Differentiation
Written by Men-Gen Tsai email: b89902089@ntu.edu.tw 1. Let f be defined for all real x, and suppose that
|f (x) − f (y)| ≤ (x − y)2 for all real x and y. Prove that f is constant.
Proof: |f (x) − f (y)| ≤ (x − y)2 for all real x and y. Fix y, |f (x)−f (y) x−y | ≤
|x − y|. Let x → y, therefore, 0 ≤ lim
x→y
f (x) − f (y) x − y ≤ lim
x→y|x − y| = 0
It implies that (f (x) − f (y))/(x − y) → 0 as x → y. Hence f0(y) = 0, f = const.
2. Suppose f0(x) > 0 in (a, b). Prove that f is strictly increasing in (a, b), and let g be its inverse function. Prove that g is differentible, and that
g0(f (x)) = 1
f0(x) (a < x < b).
Proof: For every pair x > y in (a, b), f (x) − f (y) = f0(c)(x − y) where y < c < x by Mean-Value Theorem. Note that c ∈ (a, b) and f0(x) > 0 in (a, b), hence f0(c) > 0. f (x) − f (y) > 0, f (x) > f (y) if x > y, f is strictly increasing in (a, b).
Let ∆g = g(x0+ h) − g(x0). Note that x0 = f (g(x0)), and thus, (x0+ h) − x0 = f (g(x0+ h)) − f (g(x0)),
h = f (g(x0) + ∆g) − f (g(x0)) = f (g + ∆g) − f (g).
Thus we apply the fundamental lemma of differentiation, h = [f0(g) + η(∆g)]∆g,
1
f0(g) + η(∆g) = ∆g h
Note that f0(g(x)) > 0 for all x ∈ (a, b) and η(∆g) → 0 as h → 0, thus,
h→0lim∆g/h = lim
h→0
1
f0(g) + η(∆g) = 1 f0(g(x)). Thus g0(x) = f0(g(x))1 , g0(f (x)) = f01(x).
3. Suppose g is a real function on R1, with bounded derivative (say
|g0| ≤ M ). Fix > 0, and define f (x) = x + g(x). Prove that f is one-to-one if is small enough. (A set of admissible values of can be determined which depends only on M .)
Proof: For every x < y, and x, y ∈ R, we will show that f (x) 6= f (y).
By using Mean-Value Theorem:
g(x) − g(y) = g0(c)(x − y) where x < c < y, (x − y) + ((x) − g(y)) = (g0(c) + 1)(x − y), that is,
f (x) − f (y) = (g0(c) + 1)(x − y). (∗)
Since |g0(x)| ≤ M , −M ≤ g0(x) ≤ M for all x ∈ R. Thus 1 − M ≤
g0(c) + 1 ≤ 1 + M , where x < c < y. Take c = 2M1 , and g0(c) + 1 > 0 where x < c < y for all x, y. Take into equation (*), and f (x)−f (y) < 0 since x − y < 0, that is, f (x) 6= f (y), that is, f is one-to-one (injective).
4. If
C0+C1
2 + ... + Cn−1
n + Cn n + 1 = 0,
where C0, ..., Cn are real constants, prove that the equation C0+ C1x + ... + Cn−1xn−1+ Cnxn= 0 has at least one real root between 0 and 1.
Proof: Let f (x) = C0x + ... + n+1Cn xn+1. f is differentiable in R1 and f (0) = f (1) = 0. Thus, f (1) − f (0) = f0(c) where c ∈ (0, 1) by Mean-Value Theorem. Note that
f0(x) = C0+ C1x + ... + Cn−1xn−1+ Cnxn.
Thus, c ∈ (0, 1) is one real root between 0 and 1 of that equation.
5. Suppose f is defined and differentiable for every x > 0, and f0(x) → 0 as x → +∞. Put g(x) = f (x + 1) − f (x). Prove that g(x) → 0 as x → +∞.
Proof: f (x + 1) − f (x) = f0(c)(x + 1 − x) where x < c < x + 1 by Mean-Value Theorem. Thus, g(x) = f0(c) where x < c < x + 1, that is,
x→+∞lim g(x) = lim
x→+∞f0(c) = lim
c→+∞f0(c) = 0.
6. Suppose
(a) f is continuous for x ≥ 0, (b) f0(x) exists for x > 0, (c) f (0) = 0,
(d) f0 is monotonically increasing.
Put
g(x) = f (x)
x (x > 0) and prove that g is monotonically increasing.
Proof: Our goal is to show g0(x) > 0 for all x > 0
⇔ g0(x) = xf0(x)−f (x)x2 > 0 ⇔ f0(x) > f (x)x .
Since f0(x) exists, f (x) − f (0) = f0(c)(x − 0) where 0 < c < x by Mean-Value Theorem. ⇒ f0(c) = f (x)x where 0 < c < x. Since f0 is monotonically increasing, f0(x) > f0(c), that is, f0(x) > f (x)x for all x > 0.
7. Suppose f0(x), g0(x) exist, g0(x) 6= 0, and f (x) = g(x) = 0. Prove that limt→x
f (t)
g(t) = f0(x) g0(x). (This holds also for complex functions.)
Proof:
f0(t)
g0(t) = limt→x f (t)−f (x) t−x
limt→x g(t)−g(x)t−x = limt→xf (t) limt→xg(t) = lim
t→x
f (t) g(t) Surely, this holds also for complex functions.
8. Suppose f0(x) is continuous on [a, b] and > 0. Prove that there exists δ > 0 such that
|f (t) − f (x)
t − x − f0(x)| <
whenever 0 < |t − x| < δ, a ≤ x ≤ b, a ≤ t ≤ b. (This could be expressed by saying f is uniformly differentiable on [a, b] if f0 is contin- uous on [a, b].) Does this hold for vector-valued functions too?
Proof: Since f0(x) is continuous on a compact set [a, b], f0(x) is uni- formly continuous on [a, b]. Hence, for any > 0 there exists δ > 0 such that
|f0(t) − f0(x)| <
whenever 0 < |t − x| < δ, a ≤ x ≤ b, a ≤ t ≤ b. Thus, f0(c) =
f (t)−f (x)
t−x where c between t and x by Mean-Value Theorem. Note that 0 < |c − x| < δ and thus |f0(c) − f0(x)| < , thus,
|f (t) − f (x)
t − x − f0(x)| < whenever 0 < |t − x| < δ, a ≤ x ≤ b, a ≤ t ≤ b.
Note: It does not hold for vector-valued functions. If not, take f (x) = (cos x, sin x),
[a, b] = [0, 2π], and x = 0. Hence f0(x) = (− sin x, cos x). Take any 1 > > 0, there exists δ > 0 such that
|f (t) − f (0)
t − 0 − f0(0)| <
whenever 0 < |t| < δ by our hypothesis. With calculating,
|(cos t − 1 t ,sin t
t ) − (0, 1)| <
|(cos t − 1 t ,sin t
t − 1)| < (cos t − 1
t )2 + (sin t
t − 1)2 < 2 < 2
t2 + 1 − 2(cos t + sin t) t < since 1 > > 0. Note that
2
t2 + 1 − 4 t < 2
t2 + 1 − 2(cos t + sin t) t
But t22 + 1 − 4t → +∞ as t → 0. It contradicts.
9. Let f be a continuous real function on R1, of which it is known that f0(x) exists for all x 6= 0 and that f0(x) → 0 as x → 0. Dose it follow that f0(0) exists?
Note: We prove a more general exercise as following.
Suppose that f is continuous on an open interval I containing x0, sup- pose that f0 is defined on I except possibly at x0, and suppose that f0(x) → L as x → x0. Prove that f0(x0) = L.
Proof of the Note: Using L’Hospital’s rule:
h→0lim
f (x0 + h) − f (x0)
h = lim
h→0f0(x0+ h) By our hypothesis: f0(x) → L as x → x0. Thus,
h→0lim
f (x0+ h) − f (x0)
h = L,
Thus f0(x0) exists and
f0(x0) = L.
10. Suppose f and g are complex differentiable functions on (0, 1), f (x) → 0, g(x) → 0, f0(x) → A, g0(x) → B as x → 0, where A and B are complex numbers, B 6= 0. Prove that
x→0lim f (x) g(x) = A
B. Compare with Example 5.18. Hint:
f (x)
g(x) = (f (x)
x − A) x
g(x) + A x g(x).
Apply Theorem 5.13 to the real and imaginary parts of f (x)x and g(x)x .
Proof: Write f (x) = f1(x) + if2(x), where f1(x), f2(x) are real-valued functions. Thus,
df (x)
dx = df1(x)
dx + idf2(x) dx , Apply L’Hospital’s rule to f1x(x) and f2x(x), we have
x→0lim f1(x)
x = lim
x→0f10(x)
x→0lim f2(x)
x = lim
x→0f20(x) Combine f1(x) and f2(x), we have
x→0lim f1(x)
x + i lim
x→0
f2(x)
x = lim
x→0
f1(x)
x + if2(x) x = lim
x→0
f (x) x or
x→0lim f1(x)
x + i lim
x→0
f2(x)
x = lim
x→0f10(x) + i lim
x→0f20(x) = lim
x→0f0(x) Thus, limx→0f (x)x = limx→0f0(x). Similarly, limx→0 g(x)x = limx→0g0(x).
Note that B 6= 0. Thus,
x→0lim f (x)
g(x) = lim
x→0(f (x)
x − A) x
g(x)+ A x g(x)
= (A − A)1 B + A
B = A B.
Note: In Theorem 5.13, we know g(x) → +∞ as x → 0. (f (x) = x, and g(x) = x + x2ex2i ).
11. Suppose f is defined in a neighborhood of x, and suppose f ”(x) exists.
Show that
h→0lim
f (x + h) + f (x − h) − 2f (x)
h2 = f00(x).
Show by an example that the limit may exist even if f ”(x) dose not.
Hint: Use Theorem 5.13.
Proof: By using L’Hospital’s rule: (respect to h.)
h→0lim
f (x + h) + f (x − h) − 2f (x)
h2 = lim
h→0
f0(x + h) − f0(x − h) 2h
Note that
f00(x) = 1
2(f00(x) + f00(x))
= 1
2(lim
h→0
f0(x + h) − f0(x)
h + lim
h→0
f0(x − h) − f0(x)
−h )
= 1
2lim
h→0
f0(x + h) − f0(x − h) h
= lim
h→0
f0(x + h) − f0(x − h) 2h
Thus,
f (x + h) + f (x − h) − 2f (x)
h2 → f00(x)
as h → 0. Counter-example: f (x) = x|x| for all real x.
12. If f (x) = |x|3, compute f0(x), f00(x) for all real x, and show that f(3)(0) does not exist.
Proof: f0(x) = 3|x|2 if x 6= 0. Consider f (h) − f (0)
h = |h|3 h
Note that |h|/h is bounded and |h|2 → 0 as h → 0. Thus, f0(0) = lim
h→0
f (h) − f (0)
h = 0.
Hence, f0(x) = 3|x|2 for all x. Similarly, f00(x) = 6|x|.
Thus,
f00(h) − f (0)
h = 6|h|
h
Since |h|h = 1 if h > 0 and = −1 if h < 0, f000(0) does not exist.
13. Suppose a and c are real numbers, c > 0, and f is defined on [−1, 1] by f (x) =
xasin (x−c) (if x 6= 0),
0 (if x = 0).
Prove the following statements:
(a) f is continuous if and only if a > 0.
(b) f0(0) exists if and only if a > 1.
(c) f0 is bounded if and only if a ≥ 1 + c.
(d) f0 is continuous if and only if a > 1 + c.
(e) f00(0) exists if and only if a > 2 + c.
(f) f00 is bounded if and only if a ≥ 2 + 2c.
(g) f00 is continuous if and only if a > 2 + 2c.
Proof: For (a): (⇒) f is continuous iff for any sequence {xn} → 0 with xn 6= 0, xansin x−cn → 0 as n → ∞. In particular, take
xn= ( 1
2nπ + π/2)1c > 0
and thus xan → 0 as n → ∞. Hence a > 0. (If not, then a = 0 or a < 0.
When a = 0, xan = 1. When a < 0, xan = 1/x−an → ∞ as n → ∞. It contradicts.)
(⇐) f is continuous on [−1, 1] − {0} clearly. Note that
−|xa| ≤ xasin (x−c) ≤ |xa|,
and |xa| → 0 as x → 0 since a > 0. Thus f is continuous at x = 0.
Hence f is continuous.
For (b): f0(0) exists iff xa−1sin (x−c) → 0 as x → 0. In the previous proof we know that f0(0) exists if and only if a − 1 > 0. Also, f0(0) = 0.
14. Let f be a differentiable real function defined in (a, b). Prove that f is convex if and only if f0 is monotonically increasing. Assume next that f00(x) exist for every x ∈ (a, b), and prove that f is convex if and only if f00(x) ≥ 0 for all x ∈ (a, b).
15. Suppose a ∈ R1, f is a twice-differentiable real function on (a, ∞), and M0, M1, M2 are the least upper bounds of |f (x)|, |f0(x)|, |f00(x)|, respectively, on (a, ∞). Prove that
M12 ≤ 4M0M2. Hint: If h > 0, Taylor’s theorem shows that
f0(x) = 1
2h[f (x + 2h) − f (x)] − hf00(ξ) for some ξ ∈ (x, x + 2h). Hence
|f0(x)| ≤ hM2+M0 h .
To show that M12 = 4M0M2 can actually happen, take a = −1, define
f (x) =
2x2− 1 (−1 < x < 0),
x2−1
x2+1 (0 ≤ x < ∞),
and show that M0 = 1, M1 = 4, M2 = 4. Does M12 ≤ 4M0M2 hold for vector-valued functions too?
Proof: Suppose h > 0. By using Taylor’s theorem:
f (x + h) = f (x) + hf0(x) + h2 2 f00(ξ)
for some x < ξ < x + 2h. Thus
h|f0(x)| ≤ |f (x + h)| + |f (x)| + h2
2 |f00(ξ)|
h|f0(x)| ≤ 2M0 +h2 2 M2. h2M2− 2h|f0(x)| + 4M0 ≥ 0 (∗)
Since equation (*) holds for all h > 0, its determinant must be non- positive.
4|f0(x)|2− 4M2(4M0) ≤ 0
|f0(x)|2 ≤ 4M0M2 (M1)2 ≤ 2M0M2 Note: There is a similar exercise:
Suppose f (x)(−∞ < x < +∞) is a twice-differentiable real function, and
Mk= sup
−∞<x<+∞|f(k)(x)| < +∞ (k = 0, 1, 2).
Prove that M12 ≤ 2M0M2.
Proof of Note:
f (x + h) = f (x) + f0(x)h + f00(ξ1) 2 h2
(x < ξ1 < x + h or x > ξ1 > x + h) . . . (*) f (x − h) = f (x) − f0(x)h + f00(ξ2)
2 h2
(x − h < ξ2 < x or x − h > ξ2 > x) . . . (**) (*) minus (**):
f (x + h) − f (x − h) = 2f0(x)h + h2
2(f00(ξ1) − f00(ξ2)).
2h|f0(x)| ≤ |2hf0(x)|
2h|f0(x)| ≤ |f (x + h)| + |f (x − h)| + h2
2 (|f00(ξ1)| + |f00(ξ2)|) 2h|f0(x)| ≤ 2M0+ h2M2
M2h2− 2|f0(x)|h + 2M0 ≥ 0
Since this equation holds for all h, its determinant must be non-positive:
4|f0(x)|2− 4M2(2M0) ≤ 0,
|f0(x)|2 ≤ 2M0M2 Thus
M12 ≤ 2M0M2
16. Suppose f is twice-differentiable on (0, ∞), f00 is bounded on (0, ∞), and f (x) → 0 as x → ∞. Prove that f0(x) → 0 as x → ∞. Hint: Let a → ∞ in Exercise 15.
Proof: Suppose a ∈ (0, ∞), and M0, M1, M2 are the least upper bounds of |f (x)|, |f0(x)|, |f00(x)| on (a, ∞). Hence, M12 ≤ 4M0M2. Let a → ∞, M0 = sup |f (x)| → 0. Since M2 is bounded, therefore M12 → 0 as a → ∞. It implies that sup |f0(x)| → 0 as x → ∞.
17. Suppose f is a real, three times differentiable function on [−1, 1], such that
f (−1) = 0, f (0) = 0, f (1) = 1, f0(0) = 0.
Prove that f(3)(x) ≥ 3 for some x ∈ (−1, 1).
Note that equality holds for 12(x3+ x2).
Hint: Use Theorem 5.15, with α = 0 and β = 1, −1, to show that there exist s ∈ (0, 1) and t ∈ (−1, 0) such that
f(3)(s) + f(3)(t) = 6.
Proof: By Theorem 5.15, we take α = 0, β = 1, f (1) = f (0) + f0(0) + f00(0)
2 + f(3)(s) 6 where s ∈ (0, 1). Take α = 0, andβ = −1,
f (−1) = f (0) − f0(0) + f00(0)
2 − f(3)(t) 6 where t ∈ (−1, 0). Thus
1 = f00(0)
2 +f(3)(s)
6 , s ∈ (0, 1) (∗) 0 = f00(0)
2 − f(3)(s)
6 , s ∈ (−1, 0) (∗∗) Equation (*) - equation (**):
f(3)(s)
6 + f(3)(t)
6 , s ∈ (0, 1), t ∈ (−1, 0).
f(3)(s) + f(3)(t) = 6, s, t ∈ (−1, 1).
f(3)(x) ≥ 3 f or some x ∈ (−1, 1).
Theorem 5.15: Suppose f is a real function on [a, b], n is a positive integer, f(n−1) is continuous on [a, b], f(n)(t) exists for every t ∈ (a, b).
Let α, β be distinct points of [a, b], and define P (t) =
n−1
X
k=0
f(k)(α)
k! (t − α)k.
Then there exists a point x between α and β such that f (β) = P (β) +f(n)(x)
n! (β − α)n.
18. Suppose f is a real function on [a, b], n is a positive integer, and f(n−1) exists for every t ∈ [a, b]. Let α, β, and P be as in Taylor’s theorem (5.15). Define
Q(t) = f (t) − f (β) t − β for t ∈ [a, b], t 6= β, differentiate
f (t) − f (β) = (t − β)Q(t)
n − 1 times at t = α, and derive the following version of Taylor’s theorem:
f (β) = P (β) + Q(n−1)(α)
(n − 1)! (β − α)n.
19. Suppose f is defined in (−1, 1) and f0(0) exists. Suppose −1 < αn <
βn < 1, αn→ 0, and βn→ 0 as n → ∞. Define the difference quotients Dn= f (βn) − f (αn)
βn− αn Prove the following statements:
(a) If αn < 0 < βn, then lim Dn= f0(0).
(b) If 0 < αn< βnand {βn/(βn−αn)} is bounded, then lim Dn= f0(0).
(c) If f0 is continuous in (−1, 1), then lim Dn = f0(0).
Give an example in which f is differentiable in (−1, 1) (but f0 is not continuous at 0) and in which αn, βn tend to 0 in such a way that
lim Dn exists but is different from f0(0).
Proof: For (a):
Dn= f (βn) − f (0) βn
βn
βn− αn +f (αn) − f (0) αn
−αn βn− αn Note that
f0(0) = lim
n→∞
f (αn) − f (0)
αn = lim
n→∞
f (βn) − f (0) βn Thus for any > 0, there exists N such that
L − < f (αn) − f (0)
αn < L + , L − < f (βn) − f (0)
βn < L + ,
whenever n > N where L = f0(0) respectively. Note that βn/(βn− αn) and −αn/(βn− αn) are positive. Hence,
βn
βn− αn(L − ) < f (βn) − f (0) βn
βn
βn− αn < βn
βn− αn(L + )
−αn βn− αn
(L − ) < f (αn) − f (0) αn
−αn βn− αn
< −αn βn− αn
(L + ) Combine two inequations,
L − < Dn< L + Hence, lim Dn = L = f0(0).
For (b): We process as above prove, but note that −αn/(βn− αn) < 0.
Thus we only have the following inequations:
βn
βn− αn(L − ) < f (βn) − f (0) βn
βn
βn− αn < βn
βn− αn(L + )
−αn
βn− αn(L + ) < f (αn) − f (0) αn
−αn
βn− αn < −αn
βn− αn(L − ) Combine them:
L − βn+ αn
βn− αn < Dn< L + βn+ αn βn− αn Note that {βn/(βn− αn)} is bounded, ie,
| βn βn− αn
| ≤ M
for some constant M . Thus
|βn+ αn
βn− αn| = | 2βn
βn− αn − 1| ≤ 2M + 1 Hence,
L − (2M + 1) < Dn< L + (2M + 1)
Hence, lim Dn = L = f0(0).
For (c): By using Mean-Value Theorem, Dn = f0(tn) where tn is between αn and βn. Note that
min{αn, βn} < tn< max{αn, βn} and
max{αn, βn} = 1
2(αn+ βn+ |αn− βn|) min{αn, βn} = 1
2(αn+ βn− |αn− βn|)
Thus, max{αn, βn} → 0 and min{αn, βn} → 0 as αn→ 0 and βn→ 0.
By squeezing principle for limits, tn → 0. With the continuity of f0, we have
lim Dn = lim f0(tn) = f0(lim tn) = f0(0).
Example: Let f be defined by
f (x) =
x2sin(1/x) (x 6= 0),
0 (x = 0).
Thus f0(x) is not continuous at x = 0, and f0(0) = 0. Take αn =
1
π/2+2nπ and βn= 2nπ1 . It is clear that αn→ 0, and βn→ 0 as n → ∞.
Also,
Dn = −4nπ
π(π/2 + 2nπ) → −2 π
as n → ∞. Thus, lim Dn= −2/π exists and is different from 0 = f0(0).
20.
21.
22. Suppose f is a real function on (−∞, ∞). Call x a fixed point of f if f (x) = x.
(a) If f is differentiable and f0(t) 6= 1 for every real t, prove that f has at most one fixed point.
(b) Show that the function f defined by f (t) = t + (1 + et)−1
has no fixed point, although 0 < f0(t) < 1 for all real t.
(c) However, if there is a constant A < 1 such that |f0(t)| ≤ A for all real t, prove that a fixed point x of f exists, and that x = lim xn, where x1 is an arbitrary real number and
xn+1 = f (xn) for n = 1, 2, 3, ...
(d) Show that the process describe in (c) can be visualized by the zig- zag path
(x1, x2) → (x2, x2) → (x2, x3) → (x3, x3) → (x3, x4) → ....
Proof: For (a): If not, then there exists two distinct fixed points, say x and y, of f . Thus f (x) = x and f (y) = y. Since f is differeniable, by applying Mean-Value Theorem we know that
f (x) − f (y) = f0(t)(x − y)
where t is between x and y. Since x 6= y, f0(t) = 1. It contradicts.
For (b): We show that 0 < f0(t) < 1 for all real t first:
f0(t) = 1 + (−1)(1 + et)−2et= 1 − et (1 + et)2 Since
et> 0
(1 + et)2 = (1 + et)(1 + et) > 1(1 + et) = 1 + et> et> 0 for all real t, thus
(1 + et)−2et> 0 (1 + et)−2et< 1 for all real t. Hence 0 < f0(t) < 1 for all real t.
Next, since f (t) − t = (1 − et)−1 > 0 for all real t, f (t) has no fixed point.
For (c): Suppose xn+16= xnfor all n. (If xn+1 = xn, then xn= xn+1= ... and xn is a fixed point of f ).
By Mean-Value Theorem,
f (xn+1) − f (xn) = f0(tn)(xn+1− xn)
where tn is betweem xn and xn+1. Thus,
|f (xn+1) − f (xn)| = |f0(tn)||(xn+1− xn)|
Note that |f0(tn)| is bounded by A < 1, f (xn) = xn+1, and f (xn+1) = xn+2. Thus
|xn+2− xn+1| ≤ A|xn+1− xn|
|xn+1− xn| ≤ CAn−1
where C = |x2− x1|. For two positive integers p > q,
|xp− xq| ≤ |xp − xp−1| + ... + |xq+1− xq|
= C(Aq−1+ Aq+ ... + Ap−2)
≤ CAq−1 1 − A. Hence
|xp − xq| ≤ CAq−1 1 − A.
Hence, for any > 0, there exists N = [logA(1−A)C ] + 2 such that
|xp− xq| < whenever p > q ≥ N . By Cauchy criterion we know that {xn} converges to x. Thus,
n→∞lim xn+1 = f ( lim
n→∞xn) since f is continuous. Thus,
x = f (x).
x is a fixed point of f .
For (d): Since xn+1 = f (xn), it is trivial.
23.
24.
25. Suppose f is twice differentiable on [a, b], f (a) < 0, f (b) > 0, f0(x) ≥ δ > 0, and 0 ≤ f00(x) ≤ M for all x ∈ [a, b]. Let ξ be the unique point in (a, b) at which f (ξ) = 0.
Complete the details in the following outline of Newton’s method for computing ξ.
(a) Choose x1 ∈ (ξ, b), and define {xn} by xn+1= xn− f (xn)
f0(xn).
Interpret this geometrically, in terms of a tangent to the graph of f . (b) Prove that xn+1 < xn and that
n−>oolim xn = ξ.
(c) Use Taylor’s theorem to show that xn+1− ξ = f00(tn)
2f0(xn)(xn− ξ)2 for some tn ∈ (ξ, xn).
(d) If A = M2δ, deduce that
0 ≤ xn+1− ξ ≤ 1
A[A(x1− ξ)]2n. (Compare with Exercise 16 and 18, Chap. 3)
(e) Show that Newton’s method amounts to finding a fixed point of the function g defined by
g(x) = x − f (x) f0(x). How does g0(x) behave for x near ξ?
(f) Put f (x) = x1/3 on (−∞, ∞) and try Newton’s method. What happens?
Proof: For (a): You can see the picture in the following URL:
http://archives.math.utk.edu/visual.calculus/3/newton.5/1.html.
For (b): We show that xn ≥ xn+1 ≥ ξ. (induction). By Mean-Value Theorem, f (xn) − f (ξ) = f0(cn)(xn − ξ) where cn ∈ (ξ, xn). Since f00≥ 0, f0 is increasing and thus
f (xn)
xn− ξ = f0(cn) ≤ f0(xn) = f (xn) xn− xn+1 f (xn)(xn− ξ) ≤ f (xn)(xn− xn+1)
Note that f (xn) > f (ξ) = 0 since f0 ≥ δ > 0 and f is strictly increasing.
Thus,
xn− ξ ≤ xn− xn+1 ξ ≤ xn+1
Note that f (xn) > 0 and f0(xn) > 0. Thus xn+1 < xn. Hence, xn> xn+1≥ ξ.
Thus, {xn} converges to a real number ζ. Suppose ζ 6= ξ, then xn+1 = xn− f (xn)
f0(xn)
Note that ff (x0(xnn)) > f (ζ)δ . Let α = f (ζ)δ > 0, be a constant. Thus, xn+1< xn− α
for all n. Thus, xn < x1− (n − 1)α, that is, xn → −∞ as n → ∞. It contradicts. Thus, {xn} converges to ξ.
For (c): By using Taylor’s theorem,
f (ξ) = f (xn) + f0(xn)(ξ − xn) + f00(tn) 2(xn− ξ)2 0 = f (xn) + f0(xn)(ξ − xn) + f00(tn)
2(xn− ξ)2
0 = f (xn)
f0(xn) − xn+ ξ + f00(tn)
2f0(xn)(xn− ξ)2 xn+1− ξ = f00(tn)
2f0(xn)(xn− ξ)2 where tn ∈ (ξ, xn).
For (d): By (b) we know that 0 ≤ xn+1− ξ for all n. Next by (c) we know that
xn+1− ξ = f00(tn)
2f0(xn)(xn− ξ)2 Note that f00 ≤ M and f0 ≥ δ > 0. Thus
xn+1− ξ ≤ A(xn− ξ)2 ≤ 1
A(A(x1− ξ))2n by the induction. Thus,
0 ≤ xn+1− ξ ≤ 1
A[A(x1− ξ)]2n.
For (e): If x0 is a fixed point of g(x), then g(x0) = x0, that is, x0− f (x0)
f0(x0) = x0 f (x0) = 0.
It implies that x0 = ξ and x0 is unique since f is strictly increasing.
Thus, we choose x1 ∈ (ξ, b) and apply Newton’s method, we can find out ξ. Hence we can find out x0.
Next, by calculating
g0(x) = f (x)f00(x) f0(x)2 0 ≤ g0(x) ≤ f (x)M
δ2.
As x near ξ from right hand side, g0(x) near f (ξ) = 0.
For (f ): xn+1 = xn− ff (x0(xnn)) = −2xn by calculating. Thus, xn= (−2)n−1x1
for all n, thus {xn} does not converges for any choice of x1, and we cannot find ξ such that f (ξ) = 0 in this case.
26. Suppose f is differentiable on [a, b], f (a) = 0, and there is a real number A such that |f0(x)| ≤ A|f (x)| on [a, b]. Prove that f (x) = 0 for all x ∈ [a, b]. Hint: Fix x0 ∈ [a, b], let
M0 = sup |f (x)|, M1 = sup |f0(x)|
for a ≤ x ≤ x0. For any such x,
|f (x)| ≤ M1(x0− a) ≤ A(x0− a)M0.
Hence M0 = 0 if A(x0− a) < 1. That is, f = 0 on [a, x0]. Proceed.
Proof: Suppose A > 0. (If not, then f = 0 on [a, b] clearly.) Fix x0 ∈ [a, b], let
M0 = sup |f (x)|, M1 = sup |f0(x)|
for a ≤ x ≤ x0. For any such x,
f (x) − f (a) = f0(c)(x − a)
where c is between x and a by using Mean-Value Theorem. Thus
|f (x)| ≤ M1(x − a) ≤ M1(x0 − a) ≤ A(x0− a)M0
Hence M0 = 0 if A(x0 − a) < 1. That is, f = 0 on [a, x0] by taking x0 = a + 2A1 . Repeat the above argument by replacing a with x0, and note that 2A1 is a constant. Hence, f = 0 on [a, b].
27. Let φ be a real function defined on a rectangle R in the plane, given by a ≤ x ≤ b, α ≤ y ≤ β. A solution of the initial-value problem
y0 = φ(x, y), y(a) = c (α ≤ c ≤ β)
is, by definition, a differentiable function f on [a, b] such that f (a) = c, α ≤ f (x) ≤ β, and
f0(x) = φ(x, f (x)) (a ≤ x ≤ b)
Prove that such a problem has at most one solution if there is a constant A such that
|φ(x, y2) − φ(x, y1)| ≤ A|y2− y1| whenever (x, y1) ∈ R and (x, y2) ∈ R.
Hint: Apply Exercise 26 to the difference of two solutions. Note that this uniqueness theorem does not hold for the initial-value problem
y0 = y1/2, y(0) = 0,
which has two solutions: f (x) = 0 and f (x) = x2/4. Find all other solutions.
Proof: Suppose y1 and y2 are solutions of that problem. Since
|φ(x, y2) − φ(x, y1)| ≤ A|y2− y1|,
y(a) = c, y10 = φ(x, y1), and y02 = φ(x, y2), by Exercise 26 we know that y1− y2 = 0, y1 = y2. Hence, such a problem has at most one solution.
Note: Suppose there is initial-value problem y0 = y1/2, y(0) = 0.
If y1/2 6= 0, then y1/2dy = dx. By integrating each side and noting that y(0) = 0, we know that f (x) = x2/4. With y1/2 = 0, or y = 0. All solutions of that problem are
f (x) = 0 and f (x) = x2/4
Why the uniqueness theorem does not hold for this problem? One reason is that there does not exist a constant A satisfying
|y10 − y20| ≤ A|y1− y2|
if y1 and y2 are solutions of that problem. (since 2/x → ∞ as x → 0 and thus A does not exist).
28. Formulate and prove an analogous uniqueness theorem for systems of differential equations of the form
yj0 = φj(x, y1, ..., yk), yj(a) = cj (j = 1, ..., k) Note that this can be rewritten in the form
y0 = φ(x, y), y(a) = c
where y = (y1, ..., yk) ranges over a k-cell, φ is the mapping of a (k + 1)- cell into the Euclidean k-space whose components are the function φ1, ..., φk, and c is the vector (c1, ..., ck). Use Exercise 26, for vector- valued functions.
Theorem: Let φj(j = 1, ..., k) be real functions defined on a rectangle Rj in the plane given by a ≤ x ≤ b, αj ≤ yj ≤ βj.
A solution of the initial-value problem
y0j = φ(x, yj), yj(a) = cj (αj ≤ cj ≤ βj)
is, by definition, a differentiable function fj on [a, b] such that fj(a) = cj, αj ≤ fj(x) ≤ βj, and
fj0(x) = φj(x, fj(x)) (a ≤ x ≤ b)
Then this problem has at most one solution if there is a constant A such that
|φj(x, yj2) − φj(x, yj1)| ≤ A|yj2 − yj1| whenever (x, yj1) ∈ Rj and (x, yj2) ∈ Rj.
Proof: Suppose y1 and y2 are solutions of that problem. For each components of y1 and y2, say y1j and y2j respectively, y1j = y2j by using Exercise 26. Thus, y1 = y2
29. Specialize Exercise 28 by considering the system yj0 = yj+1 (j = 1, ..., k − 1),
y0k= f (x) −
k
X
j=1
gj(x)yj
where f, g1, ..., gk are continuous real functions on [a, b], and derive a uniqueness theorem for solutions of the equation
y(k)+ gk(x)y(k−1)+ ... + g2(x)y0+ g1(x)y = f (x), subject to initial conditions
y(a) = c1, y0(a) = c1, ..., y(k−1)(a) = ck.
Theorem: Let Rj be a rectangle in the plain, given by a ≤ x ≤ b, min yj ≤ yj ≤ max yj. (since yj is continuous on the compact set, say
[a, b], we know that yj attains minimal and maximal.) If there is a constant A such that
|yj+1,1− yj+1,2| ≤ A|yj,1− yj,2| (j < k)
|Pkj=1gj(x)(yj,1− yj,2)| ≤ A|yk,1− yk,2| whenever (x, yj,1) ∈ Rj and (x, yj,2) ∈ Rj.
Proof: Since the system y10, ..., yk0 with initial conditions satisfies a fact that there is a constant A such that |y10 −y20| ≤ A|y1−y2|, that system has at most one solution. Hence,
y(k)+ gk(x)y(k−1)+ ... + g2(x)y0+ g1(x)y = f (x), with initial conditions has at most one solution.