You are on page 1of 3

Hints to Exercise 5.

3
(1) (a) (b) (c) (d) True. By the Fundamental Theorem Of Calculus and Chain Rule. True. By Integration By Parts. True. By Integration By Parts. False. g(f (a)) or g(f (b)) may be undened. However, if they are dened and g(f (a)), g(f (b)) [a, b], then the assertion holds by applying Change Of Variables repeatedly. F (x) = 1x f (t)dt, hence F (x) = f (x2 )(x2 ) = 2xf (x2 ). 3 2 F (x) = 0x f (t)dt 0x f (t)dt, hence F (x) = 3x2 f (x3 ) 2xf (x2 ). F (x) = (x cos xf (x cos x))(x cos x) = (x cos2 xx2 sin x cos x)f (x cos x). Apply Change of Variables and use the Leibnizian notation. 0 F (x) = 0x f (t x)d(t x) = x f (t)dt = 0x f (t)dt.
2

(2) (a) (b) (c) (d)

(3) Use Change of Variables and the Comparison Theorem. (4) Use Change of Variables and the Comparison Theorem. Note that x3 = ((x2 + 1) 1)3/2 and x2 = 1 ( 1 x2 )2 . (5) Use Integration by Parts. (6) Since f C 1 [a, b], f is continuous on [a, b], hence integrable on [a, b], and b a f (x)dx = f (b) f (a) by the Fundamental Theorem of Calculus. On b the other hand, since f is continuous on [a, b], a f (x)dx = f (x0 )(b a) for some x0 [a, b] by the First Mean Value Theorem for Integrals. Thereb fore, f (b) f (a) = a f (x)dx = (b a)f (x0 ), where x0 [a, b].
x b (7) Let F (x) = a f (x)dx + x f (x)dx, x [a, b], hence F is zero on (a, b). Moreover, F C 1 [a, b] by the Fundamental Theorem of Calculus, so F must be zero at x = a and x = b, hence on [a,b]. Again, by the Fundamental Theorem of Calculus, for each x [a, b] F (x) = f (x) f (x) = 0, that is, ( )f (x) = 0, equivalently, f (x) = 0 since = . a (8) (a) Apply the Fundamental Theorem of Calculus and that a f (x)dx = 0. kn k (b) Consider that L(2n ) = n 22k1 dt and L(2n ) = n 22kn1 dt . k=1 k=1 t t (c) Use Change of Variables and note that 1 t1/q x when t takes values on [1, xq ]. xy (d) Use Change of Variables and write 1xy dt = 1x dt + x dt . t t t (e) L(e) = limn L((1+1/n)n ) = limn L(1+1/n) = limn L (1+1/n)(1/n) = 1/n (1/n) 1 limn 1+1/n = 1.

(9) (a) Since L (x) = 1/x is always positive on (0, ), L is strictly increasing and continuous on (0, ) and L is never zero on (0, ). By the Inverse Function Theorem E = L1 exists as a function from R onto (0, ). Moreover E (a) = 1/L (E(a)) = E(a) for each a R, hence E (x) = E(x). Finally, By Exercise (8), L(1) = 0 and L(e) = 1, we obtain E(0) = 1 and E(1) = e. (b) By (8).b., L(0, ) = R, hence E(R) = (0, ), and so for any M > 0, there is some xM R such that E(xM ) = M . Since E is strictly increasing(by the Inverse Function Theorem), we have E(x) M whenever x xM . Thus, E(x) as x . A similar argument shows that E(x) 0 as x . (c) Set L(y) = x, which is equivalent to E(x) = y. Then by (8).c., E(xq) = E(qL(y)) = E(L(y q )) = y q = E(x)q . Hence E(q) = E(1q) = E(1)q = eq . (d) Using (8).d.,we have E(x + y) = E(L(E(x)) + L(E(y))) = E(L(E(x)E(y))) = E(x)E(y). (e) For > 0, and 0 < x < y, we have L(x) < L(y) since L is strictly increasing. Again, E is also strictly increasing, we obtain E(L(x)) < E(L(y)). By the denition of x , the last inequality means that x < y . A similar argument shows that x > y if < 0. By d., x+ = E(( + )L(x)) = E(L(x) + L(x)) = E(L(x))E(L(x)) = x x . This equality and that E(0) = 1 imply x = x1 . Finally, (x ) = E (L(x)) = E(L(x))L (x) = x (1/x) = x1 . (10) Use the Inverse Function Theorem, and write
f (b) f (b) f (a)

f 1 (x)dx
b a

b = f (a) f 1 (x)f (f 1 (x))d(f 1 (x)) = a xf (x)dx = xf (x)|b a Note that we also use change of variables here.

f (x)dx.

(11) By Theorem 5.34, f | | is integrable on [a, b]. On the other hand, since C 1 and is never zero on [a, b], 1/| | is integrable on [a, b]. Then the proof completes by Theorem 5.23. (The product of two integrable functions is still integrable) One can also directly prove the assertion by consider a partition P on [a, b] and the partition (P ) on [c, d], imitating the proof for Theorem 5.34. (12) First use Exercise 4.2.7.a to show that f q is integrable on [a,b], using Denition 5.9.

(13) First show that log an is a Riemann sum of log x with partition P = 1, 1 + 2 n1 1 ,1 + ,...,1 + ,2 n n n

generated by samples tj = 1 + j/n. Consequently show that


2 n

lim log an =

4 log x dx = log . e

Finally, since log x is continuous and monotone, we obtain 4 lim an = . n e

You might also like