Let f : [0, 1] → R be a continuous function such that for any x , y ∈ [0,1] ,
10.2.3. Let f : [0, 1] → R be a continuous function such that for any x , y ∈ [0,1] ,
xf (y) + y f (x) ≤ 1.
(i) Show that
f (x)dx ≤ .
(ii) Find a function satisfying the condition for which there is equality.
376 10 Applications of the Integral Calculus Solution. (i) We observe that
f (x)dx =
f (sin θ ) cos θ d θ =
f (cos θ ) sin θ d θ .
So, by hypothesis,
2 f (x)dx ≤
1d θ = .
√ (ii) Let f (x) = 1 −x 2 . If x = sin θ and y = sin Φ , then
xf (y) + y f (x) = sin θ cos Φ + sin Φ cos θ = sin( θ + Φ ) ≤ 1. # √
On the other hand, 1 1 −x 0 2 dx = π /4. ⊓ ⊔
10.2.4. Let # f : [a, b]→R be a function of class C 1 such that b
a f (t)dt = 0 . (i) For any integer n ≥1 define the function f n (t) = cos(2 π nt ) . Prove that f n
satisfies the above hypotheses and compute
sup 0 |f n (t)| # dt
0 |f n ′ (t)| dt
n ≥1 1 2
(ii) Show that
f 2 2 (t)dt ≤ (b − a) b (f ′ (t)) 2 dt .
be a function of class C 1 such that g (a) = g(b) = 0 (in this case, # b
(iii) Let g : [a, b]→R
a g (t)dt is not necessarily zero). Prove that
" b (b − a) 2 "
(t)dt ≤
(g ′ (t)) 2 dt .
Solution. (i) The required value is 4 π −1 . (ii) Since f is continuous and # b
a f (t)dt =0, there exists c ∈ (a,b) such that f (c)=0. Hence
f (t) = f (c) +
f ′ (u)du =
f ′ (u)du, ∀t ∈ [a,b].
By the Cauchy–Schwarz inequality we obtain for all t ∈ [a,b], " t
1 /2 | f (t)| ≤ |
f ′ (u)du| ≤ |t − c|
(f ′ (u)) 2 du .
Therefore
| f (t)| 2 dt ≤
(f ′ (u)) 2 du
|t − c|dt.
10.2 Integral Inequalities 377 To evaluate the second term, we observe that # b |t −c| ≤ b−a, so
a |t −c|dt ≤ (b−a) 2 , and the proof is concluded. An alternative proof is based on the remark that
2 +b
|t − c|dt = c − c(a + b) +
This is a quadratic function in c, whose minimum is achieved at c = (a + b)/2. The maximum is attained either at a or at b, and this value equals (b − a) 2 /2. We conclude that
2 dt
(b − a)
| f (t)|
(f ′ (u)) 2 du .
(iii) We intend to apply the above result for f =g ′ . In this case, we would find # b
a f (t)dt = g(b) − g(a) = 0. However, this in not a good idea, since it does not yield our conclusion. We remark that the points a and b play the same role
as c. This implies that it is better to restart the proof, but considering the differ- ent cases t ≤ (b + a)/2 and t ≥ (b + a)/2.
If t ≤ (b + a)/2 we have
f (t) =
f ′ (u)du,
so
| f (t)| 2 ≤ (t − a) (f ′ (u)) 2 du ≤ (t − a)
" (b+a)/2
(f ′ (u)) 2 du .
By integration we obtain " (b+a)/2
2 (b − a) " (b+a)/2
| f (t)| 2 dt ≤
(f ′ (t)) 2 dt .
If t ≥ (b + a)/2, we obtain
| f (t)| 2 dt
(b − a)
(f ′ (t)) 2 dt .
(b+a)/2
8 (b+a)/2
Adding the above two inequalities, we obtain
" b 2 (b − a) " b
| f (t)| 2 dt ≤
(f ′ (t)) 2 dt ,
which concludes the proof. ⊓ ⊔ We give in what follows an elementary proof of the Young inequality in a partic-
ular case.
10.2.5. Let f : [0, +∞)→R
be a strictly increasing function with a continuous derivative such that f (0) = 0 . Let f −1 denote the inverse of f . Prove that for all
positive numbers a and b , with b < f (a) ,
ab <
f (x)dx +
f −1 (y)dy .
378 10 Applications of the Integral Calculus Solution. We first observe that if f is a strictly increasing function with a
continuous derivative, then integration by parts yields
f (x) dx = b f (b) − a f (a) − xf ′ (x) dx.
Let y = f (x), x = f −1 (y). We obtain
" b " f (b)
f (x) dx = b f (b) − a f (a) −
f −1 (y) dy. (10.2)
a f (a)
Now if u = f (a), v = f (b), relation (10.2) becomes " f −1 (v)
Thus, we can always express # f −1 (y) dy in terms of f (x) dx. Returning to our problem, assume that 0 < r < a. Then
f (x)dx < r f (r) −
f (x) dx .
Applying relation (10.2) to # r
0 f (x)dx, we obtain
" a " f (r)
af (r) −
f (x)dx <
f −1 (y) dy.
Since 0 < b < f (a), we can take r = f −1 (b), and our conclusion follows. ⊓ ⊔