0% found this document useful (0 votes)
10 views4 pages

sol3

The document contains solutions to various problems in a mathematical analysis course, focusing on continuity, differentiability, and the properties of functions in multiple dimensions. Key results include proving the continuity of functions under certain conditions, exploring the relationships between partial derivatives, and demonstrating the differentiability of specific functions. The solutions utilize concepts such as the mean value theorem and the properties of mixed derivatives.
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)
10 views4 pages

sol3

The document contains solutions to various problems in a mathematical analysis course, focusing on continuity, differentiability, and the properties of functions in multiple dimensions. Key results include proving the continuity of functions under certain conditions, exploring the relationships between partial derivatives, and demonstrating the differentiability of specific functions. The solutions utilize concepts such as the mean value theorem and the properties of mixed derivatives.
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
You are on page 1/ 4

MATH 404 ANALYSIS II - SPRING 2009

SOLUTIONS to HOMEWORK 3

Problem 1. Show that if U ⊂ Rn and f : U → R, and if the partial derivatives


Dj f exist and are bounded on Bε (a) ⊂ U , then f is continuous at a.
Solution: Assume |Dj f (x)| < C for all x ∈ Bε (a) and 1 ≤ j ≤ n. Let h ∈ Rn
j
be such that |h| < ε. Define k0 = 0 and kj = i=1 hj ej for 1 ≤ j ≤ n. Then
a + +kj ∈ Bε (a). Then
 
 n 
 

|f (a + h) − f (a)| =  (f (a + kj ) − f (a + kj−1 )
j=1 
(1)
 n
≤ |f (a + kj−1 + hj ej ) − f (a + kj−1 )| .
j=1

If hj = 0, then f (a + kj−1 + hj ej ) − f (a + kj−1 ) = 0. For any 1 ≤ j ≤ n such


that hj = 0, let gj (t) = f (a + kj−1 + hj ej ) − f (a + kj−1 for t between 0 and
hj . Since Dj exists on Bε (a), gj is continuous on [0, hj ] if hj > 0 (or on [hj , 0] if
hj < 0) differentiable on (0, hj ) if hj > 0 (or on (hj , 0) if hj < 0). By the mean
value theorem, there is tj between 0 and hj so that gj (hj ) − g(0) = g  (tj )hj =
Dj (a + kj−1 + tj ej )hj . Abbreviating a∗j = a + kj−1 + tj ej , the right hand side of
(1) is equal to
n
 n n
     
Dj (a∗j )hj  = Dj (a∗j ) · |hj | ≤ C |hj | ≤ Cn |h| .
j=1 j=1 j=1

Consequently, f is continuous a.

Problem 2. Let f : R3 → R and g : R2 → R be differentiable. Let

F : R2 → R, F (x, y) = f (x, y, g(x, y)).

(a) Find DF in terms of the partial derivatives of f and g.


(b) If F (x, y) = 0 for all (x, y), find D1 g and D2 g in terms of partial derivatives
of f .
Solution: Denote by G : R2 → R3 the map defined by G(x, y) = (x, y, g(x, y)).
Then G is differentiable at each point (x, y) ∈ R2 (a) Note that F = f ◦G : R2 → R.
By the chain rule, DF (x, y) = Df (G(x, y))DG(x, y). That is, if p = (x, y) and
q = G(p) = (x, y, g(x, y)), then
 
    1 0
DF (p) = D1 F (p) D2 F (p) = D1 f (q) D2 f (q) D3 f (q)  0 1 
D1 g(p) D2 g(p)

 
= D1 f (q) + D3 f (q)D1 g(p) D2 f (q) + D3 f (q)D2 g(p) .
1
2

 
(b) If F (x, y) = 0 for all (x, y), then DF (x, y) = 0 0 , and, in view of the above
identity,
0 = D1 f (q) + D3 f (q)D1 g(p) and 0 = D2 f (q) + D3 f (q)D2 g(p).
Hence
D1 f (x, y, g(x, y))) D2 f (x, y, g(x, y)))
D1 g(x, y) = − and D2 g(x, y) = − .
D3 f (x, y, g(x, y)) D3 f (x, y, g(x, y))
Problem 3.
(a) Let f : R → R be defined by
x2 sin(1/x) if x = 0,
f (x) =
0 if x = 0.
Show that f is differentiable at 0 but f  is not continuous at 0.
(b) Let g : R2 → R be defined by
1
(x2 + y 2 ) sin( √ ) if (x, y) = 0,
g(x, y) = x2 +y 2
0 if (x, y) = 0.
Show that g is differentiable at (0, 0) but Di g is not continuous at (0, 0).
Solution: (a) For x = 0, f  (x) = 2x sin(1/x) − cos(1/x) and at x = 0,
x2 sin(1/x)
f  (0) = lim = lim x sin(1/x) = 0.
x→0 x x→0

For xj = 1/(2πj) for j ≥ 1, we have xj → 0 and f  (xj ) = − cos(2πj) = −1. Hence


f  (xj ) → f  (0) so that f  is not continuous.
(b) We have
g(t, 0)
D1 g(0, 0) = lim = lim t sin(1/t) = 0
t→0 t t→0
g(0t)
D2 g(0, 0) = lim = lim t sin(1/t) = 0
t→0 t t→0

and for (x, y) = (0, 0),


x
D1 g(x, y) = 2x sin(1/ x2 + y 2 ) − cos(1/ x2 + y 2 )
x2 + y 2
y
D2 g(x, y) = 2y sin(1/ x2 + y 2 ) − cos(1/ x2 + y 2 ).
x + y2
2

√ √
Take (xj , yj ) = (1/(2 2πj), 1/(2 2πj). Then (xj , yj ) → (0, 0) and x2j + yj2 =
1/(2πj). So,
1 1
D1 f (xj , yj ) = − √ cos(2πj) = − √
2 2
1 1
D2 f (xj , yj ) = − √ cos(2πj) = − √ .
2 2
Consequently, D1 f and D2 f are not continuous at (0, 0). On the othere hand, g is
differentiable at (0, 0), then Dg(0, 0) = [0 0]. We calculate the limit
g(x, y) − g(0, 0) − 0 · x − 0 · y 1
lim = lim x2 + y 2 sin( ) = 0,
(x,y)→(0,0) x2 + y 2 (x,y)→(0,0) x2 + y 2
3

showing that g is differentiable at (0, 0).


Problem 4. Define f : R2 → R by
x2 arctan( yx ) − y 2 arctan( xy ) if xy = 0,
f (x) =
0 if xy = 0.
Show that D1 D2 f (0) = D2 D1 f (0).
Solution:
f (x, y) − f (0, y)
D1 f (0, y) = lim
x→0 x
y2
limx→0 x arctan xy − x arctan xy y = 0 −1 y = 0
= =
0 y=0 0 y=0
and similarly,
1 x = 0
D2 f (x, 0) =
0 x = 0.
Hence
D1 f (0, y) − D1 f (0, 0)
D2 D1 f (0, 0) = lim = −1
y→0 y
D2 f (x, 0) − D2 f (0, 0)
D1 D2 f (0, 0) = lim = 1.
x→0 y
Problem 5. Let g : R → R be of class C 2 . Show that
g(a + h) − 2g(a) + g(a − h)
lim = g  (a).
h→0 h2
Hint: Let f (x, y) = g(x+y). If h > 0 is small, let D(h) = [f (a/2+h/2, a/2+h/2)−
f (a/2 − h/2, a/2 + h/2)]− [f (a/2 +h/2, a/2 −h/2)− f (a/2− h/2, a/2− h/2)]. Argue
like in the proof of the theorem about mixed derivatives to show that D(h)/h2 →
D2 D1 f (a/2, a/2).
Solution: Let f (x, y) = g(x + y). Since (x, y) → x + y is of class C ∞ and g is of
class C 2 , the map f is of class C 2 .
D(h) := [f (a/2 + h/2, a/2 + h/2) − f (a/2 − h/2, a/2 + h/2)]
= −[f (a/2 + h/2, a/2 − h/2) − f (a/2 − h/2, a/2 − h/2)]
and
r(h) := f (a/2 + h/2, a/2 + t) − f (a/2 − h/2, a/2 + t).
So,
D(h) = r(h/2) − r(−h/2).
The function g is continuous on [−h2, h/2] and is differentiable on (−h/2, h/2). So,
applying the mean value theorem, there is s∗ ∈ (−h/2, h/2) so that
(2) D(h) = r (s∗ )h = [D2 f (a/2 + h/2, a/2 + s∗ ) − D2 f (a/2 − h/2, a/2 + s∗ )]h.
The function t → D2 f (a/2 + t, a/2 + s∗ ) is continuous on −h/2, h/2] and differ-
entiable on (−h/2, h/2). So, applying again the mean value theorem , there is
t∗ ∈ (−h/2, h/2) for which the right side of (2) is equal to
D1 D2 f (a/2 + t∗ , a/2 + s∗ )h2 .
4

Hence D(h) = D1 D2 f (a/2 + t∗ , a/2 + s∗ )h2 . On the other hand, D(h) = g(a + h) −
2g(a) + g(a − h) and D1 D2 f (a/2 + t∗ , a/2 + s∗ ) = g  (a + s∗ + t∗ ). Moreover, if
h → 0, then s∗ , t∗ → 0 so that
g(a + h) − 2g(a) + g(a − h) D(h)
= = D1 D2 f (a/2 + t∗ , a/2 + s∗ )
h2 h2
= g  (a + s∗ + t∗ ) → g  (a)
since g  is continuous.

You might also like