156 3 Real Analysis
3.2.10 Inequalities for Integrals...................................
A very simple inequality states that iff :[a, b]→Ris a nonnegative continuous
function, then
∫b
a
f(x)dx≥ 0 ,
with equality if and only iffis identically equal to zero. Easy as this inequality looks,
its applications are often tricky. This is the case with a problem from the 1982 Romanian
Mathematical Olympiad, proposed by the second author of the book.
Example.Find all continuous functionsf:[ 0 , 1 ]→Rsatisfying
∫ 1
0
f(x)dx=
1
3
+
∫ 1
0
f^2 (x^2 )dx.
Solution.First, we would like the functions in both integrals to have the same variable.
A substitution in the first integral changes it to
∫ 1
0 f(x
(^2) ) 2 xdx. Next, we would like to
express the number^13 as an integral, and it is natural to choose
∫ 1
0 x
(^2) dx. The condition
from the statement becomes
∫ 1
0
2 xf (x^2 )dx=
∫ 1
0
x^2 +
∫ 1
0
f^2 (x^2 )dx.
This is the same as
∫ 1
0
[f^2 (x^2 )− 2 xf (x^2 )+x^2 ]dx= 0.
Note that the function under the integral,f^2 (x^2 )− 2 xf (x^2 )+x^2 =(f (x^2 )−x)^2 ,isa
perfect square, so it is nonnegative. Therefore, its integral on[ 0 , 1 ]is nonnegative, and
it can equal zero only if the function itself is identically zero. We find thatf(x^2 )=x.
Sof(x)=
√
xis the unique function satisfying the condition from the statement.
473.Determine the continuous functionsf:[ 0 , 1 ]→Rthat satisfy
∫ 1
0
f (x)(x−f (x))dx=
1
12
.
474.Letnbe an odd integer greater than 1. Determine all continuous functionsf :
[ 0 , 1 ]→Rsuch that
∫ 1
0
(
f
(
x
(^1) k))n−k
dx=
k
n
,k= 1 , 2 ,...,n− 1.