0% found this document useful (0 votes)
134 views7 pages

3 Variable

1. The document presents an inequality relating three variables a, b, c where a + b + c = 1. It uses the AM-GM inequality to express ab + bc + ca in terms of a parameter q ≥ 0. 2. The maximum and minimum values of abc are found in terms of q by considering a cubic polynomial and its zeros. This leads to a double inequality relating abc to q. 3. Several applications of the inequality are given, including solving inequality problems involving sums and products of a, b, c. Equality conditions in the problems are also characterized.
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)
134 views7 pages

3 Variable

1. The document presents an inequality relating three variables a, b, c where a + b + c = 1. It uses the AM-GM inequality to express ab + bc + ca in terms of a parameter q ≥ 0. 2. The maximum and minimum values of abc are found in terms of q by considering a cubic polynomial and its zeros. This leads to a double inequality relating abc to q. 3. Several applications of the inequality are given, including solving inequality problems involving sums and products of a, b, c. Equality conditions in the problems are also characterized.
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/ 7

On a class of three-variable inequalities

Vo Quoc Ba Can
1 Theorem
Let a, b, c be real numbers satisfying a+b +c = 1. By the AM - GM inequality, we have ab +bc +ca
1
3
,
therefore setting ab + bc + ca =
1q
2
3
(q 0), we will nd the maximum and minimum values of abc in
terms of q.
If q = 0, then a = b = c =
1
3
, therefore abc =
1
27
. If q = 0, then (ab)
2
+(b c)
2
+(c a)
2
> 0. Consider
the function f(x) = (x a)(x b)(x c) = x
3
x
2
+
1q
2
3
x abc. We have
f

(x) = 3x
2
2x +
1 q
2
3
whose zeros are x
1
=
1 + q
3
, and x
2
=
1 q
3
.
We can see that f

(x) < 0 for x


2
< x < x
1
and f

(x) > 0 for x < x


2
or x > x
1
. Furthermore, f(x) has
three zeros: a, b, and c. Then
f
_
1 q
3
_
=
(1 q)
2
(1 + 2q)
27
abc 0
and
f
_
1 + q
3
_
=
(1 + q)
2
(1 2q)
27
abc 0.
Hence
(1 + q)
2
(1 2q)
27
abc
(1 q)
2
(1 + 2q)
27
and we obtain
Theorem 1.1 If a, b, c are arbitrary real numbers such that a + b + c = 1, then setting
ab + bc + ca =
1q
2
3
(q 0), the following inequality holds
(1 + q)
2
(1 2q)
27
abc
(1 q)
2
(1 + 2q)
27
.
Or, more general,
Theorem 1.2 If a, b, c are arbitrary real numbers such that a + b + c = p, then setting
ab + bc + ca =
p
2
q
2
3
(q 0) and r = abc, we have
(p + q)
2
(p 2q)
27
r
(p q)
2
(p + 2q)
27
.
This is a powerful tool since the equality holds if and only if (a b)(b c)(c a) = 0.
Mathematical Reflections 2 (2007) 1
Here are some identities which we can use with this theorem
a
2
+ b
2
+ c
2
=
p
2
+ 2q
2
3
a
3
+ b
3
+ c
3
= pq
2
+ 3r
ab(a + b) + bc(b + c) + ca(c + a) =
p(p
2
q
2
)
3
3r
(a + b)(b + c)(c + a) =
p(p
2
q
2
)
3
r
a
2
b
2
+ b
2
c
2
+ c
2
a
2
=
(p
2
q
2
)
2
9
2pr
ab(a
2
+ b
2
) + bc(b
2
+ c
2
) + ca(c
2
+ a
2
) =
(p
2
+ 2q
2
)(p
2
q
2
)
9
pr
a
4
+ b
4
+ c
4
=
p
4
+ 8p
2
q
2
+ 2q
4
9
+ 4pr
2 Applications
2.1 Let a, b, c be positive real numbers such that a + b + c = 1. Prove that
1
a
+
1
b
+
1
c
+ 48(ab + bc + ca) 25.
Solution. We can easily check that q [0, 1], by using the theorem we have
LHS =
1 q
2
3r
+ 16(1 q
2
)
9(1 + q)
(1 q)(1 + 2q)
+ 16(1 q
2
) =
2q
2
(4q 1)
2
(1 q)(1 + 2q)
+ 25 25.
The inequality is proved. Equality holds if and only if a = b = c =
1
3
or a =
1
2
, b = c =
1
4
and their
permutations.
2.2 [Vietnam 2002] Let a, b, c be real numbers such that a
2
+ b
2
+ c
2
= 9. Prove that
2(a + b + c) abc 10.
Solution. The condition can be rewritten as p
2
+ 2q
2
= 9. Using our theorem, we have
LHS = 2p r 2p
(p + q)
2
(p 2q)
27
=
p(5q
2
+ 27) + 2q
3
27
.
We need to prove that
p(5q
2
+ 27) 270 2q
3
.
This follows from
(270 2q
3
)
2
p
2
(5q
2
+ 27)
2
,
or, equivalently,
27(q 3)
2
(2q
4
+ 12q
3
+ 49q
2
+ 146q + 219) 0.
The inequality is proved. Equality holds if and only if a = b = 2, c = 1 and their permutations.
2.3 [Vo Quoc Ba Can] For all positive real numbers a, b, c, we have
a + b
c
+
b + c
a
+
c + a
b
+ 11
_
ab + bc + ca
a
2
+ b
2
+ c
2
17.
Mathematical Reflections 2 (2007) 2
Solution. Because the inequality is homogeneous, without loss of generality, we may assume that p = 1.
Then q [0, 1] and the inequality can be rewritten as
1 q
2
3r
+ 11

1 q
2
1 + 2q
2
20.
Using our theorem, it suces to prove
11

1 q
2
1 + 2q
2
20
9(1 + q)
(1 q)(1 + 2q)
=
40q
2
+ 11 + 11
(1 q)(1 + 2q)
.
If 40q
2
+ 11q + 11 0, or q
11+3

209
80
, it is trivial. If q
11+3

209
80
<
2
3
, we have
121(1 q
2
)
(1 + 2q
2
)

(40q
2
+ 11q + 11)
2
(1 q)
2
(1 + 2q)
2
=
3q
2
(11 110q + 255q
2
+ 748q
3
1228q
4
)
(1 + 2q
2
)(1 q)
2
(1 + 2q)
2
.
On the other hand,
11 110q + 255q
2
+ 748q
3
1228q
4
= q
4
_
11
q
4

110
q
3
+
255
q
2
+
748
q
1228
_
q
4
_
11
(2/3)
4

110
(2/3)
3
+
255
(2/3)
2
+
748
2/3
1228
_
=
2435
16
q
4
0.
The inequality is proved. Equality occurs if and only if a = b = c.
2.4 [Vietnam TST 1996] Prove that for any a, b, c R, the following inequality holds
(a + b)
4
+ (b + c)
4
+ (c + a)
4

4
7
(a
4
+ b
4
+ c
4
).
Solution. If p = 0 the inequality is trivial, so we will consider the case p = 0. Without loss of generality,
we may assume p = 1. The inequality becomes
3q
4
+ 4q
2
+ 10 108r 0
Using our theorem, we have
3q
4
+ 4q
2
+ 10 108r 3q
4
+ 4q
2
+ 10 4(1 q)
2
(1 + 2q) = q
2
(q 4)
2
+ 2q
4
+ 6 0.
The inequality is proved. Equality holds only for a = b = c = 0.
2.5 [Pham Huu Duc, MR1/2007] Prove that for any positive real numbers a, b, and c,
_
b + c
a
+
_
c + a
b
+
_
a + b
c

6
a + b + c
3

abc
Solution. By Holders inequality, we have
_

cyc
_
b + c
a
_
2
_

cyc
1
a
2
(b + c)
_

cyc
1
a
_
3
It suces to prove that
_

cyc
1
a
_
3

6(a + b + c)
3

abc

cyc
1
a
2
(b + c)
Mathematical Reflections 2 (2007) 3
Setting x =
1
a
, y =
1
b
, z =
1
c
, the inequality becomes
(x + y + z)
3
6
3

xyz(xy + yz + zx)

cyc
x
y + z
,
or
(x + y + z)
3

6
3

xyz(xy + yz + zx)((x + y + z)
3
2(x + y + z)(xy + yz + zx) + 3xyz)
(x + y)(y + z)(z + x)
.
By the AM - GM inequality,
(x + y)(y + z)(z + x) = (x + y + z)(xy + yz + zx) xyz
8
9
(x + y + z)(xy + yz + zx).
It remains to prove that
4(x + y + z)
4
27
3

xyz((x + y + z)
3
2(x + y + z)(xy + yz + zx) + 3xyz).
Setting p = x + y + z, xy + yz + zx =
p
2
q
2
3
(p q 0), the inequality becomes
4p
4
9
3

xyz(p
3
+ 2pq
2
+ 9xyz).
Applying our theorem, it suces to prove that
4p
4
9
3
_
(p q)
2
(p + 2q)
27
_
p
3
+ 2pq
2
+
(p q)
2
(p + 2q)
3
_
,
4p
4

3
_
(p q)
2
(p + 2q)(3p
3
+ 6pq
2
+ (p q)
2
(p + 2q)).
Setting u =
3
_
pq
p+2q
1, the inequality is equivalent to
4(2u
3
+ 1)
4
27u
2
(4u
9
+ 5u
6
+ 2u
3
+ 1),
or
f(u) =
(2u
3
+ 1)
4
u
2
(4u
9
+ 5u
6
+ 2u
3
+ 1)

27
4
We have
f

(u) =
2(2u
3
+ 1)
3
(u
3
1)(2u
3
1)(2u
6
+ 2u
3
1)
u
3
(u
3
+ 1)
2
(4u
6
+ u
3
+ 1)
2
f

(u) = 0 u =
3

3 1
2
, or u =
1
3

3
, or u = 1.
Now, we can easily verify that
f(u) min
_
_
_
f
_
_
3

3 1
2
_
_
, f(1)
_
_
_
=
27
4
,
which is true. The inequality is proved. Equality holds if and only if a = b = c.
2.6 [Darij Grinberg] If a, b, c 0, then
a
2
+ b
2
+ c
2
+ 2abc + 1 2(ab + bc + ca).
Mathematical Reflections 2 (2007) 4
Solution. Rewrite the inequality as
6r + 3 + 4q
2
p
2
0.
If 2q p, it is trivial. If p 2q, using the theorem, it suces to prove that
2(p 2q)(p + q)
2
9
+ 3 + 4q
2
p
2
0,
or
(p 3)
2
(2p + 3) 2q
2
(2q + 3p 18).
If 2p 9, we have 2q + 3p 4p 18, therefore the inequality is true. If 2p 9, we have
2q
2
(2q + 3p 18) 4q
2
(2p 9) p
2
(2p 9) = (p 3)
2
(2p + 3) 27 < (p 3)
2
(2p + 3).
The inequality is proved. Equality holds if and only if a = b = c = 1.
2.7 [Schurs inequality] For any nonnegative real numbers a, b, c,
a
3
+ b
3
+ c
3
+ 3abc ab(a + b) + bc(b + c) + ca(c + a).
Solution. Because the inequality is homogeneous, we can assume that a +b +c = 1. Then q [0, 1] and
the inequality is equivalent to
27r + 4q
2
1 0.
If q
1
2
, it is trivial. If q
1
2
, by the theorem we need to prove that
(1 + q)
2
(1 2q) + 4q
2
1 0,
or
q
2
(1 2q) 0,
which is true. Equality holds if and only if a = b = c or a = b, c = 0 and their permutations.
2.8 [Pham Kim Hung] Find the greatest constant k such that the following inequality holds for any
positive real numbers a, b, c
a
3
+ b
3
+ c
3
(a + b)(b + c)(c + a)
+
k(ab + bc + ca)
(a + b + c)
2

3
8
+
k
3
.
Solution. For a = b = 1 +

3 and c = 1, we obtain k
9(3+2

3)
8
= k
0
. We will prove that this is the
desired value. Let k
0
be a constant satisfying the given inequality. Without loss of generality, assume
that p = 1. Then q [0, 1] and the inequality becomes
3(3r + q
2
)
3r + 1 q
2
+
k
0
(1 q
2
)
3

3
8
+
k
0
3
.
It is not dicult to verify that this is an increasing function in terms of r. If 2q 1, we have
V T
3q
2
1 q
2
+
k
0
(1 q
2
)
3
1 +
k
0
4

3
8
+
k
0
3
.
(since this is an increasing function in terms of q
2

1
4
)
If 2q 1, using our theorem, it suces to prove that
3((1 + q)
2
(1 2q) + 9q
2
)
(1 + q)
2
(1 2q) + 9(1 q
2
)
+
k
0
(1 q
2
)
3

3
8
+
k
0
3
.
We have
LHS RHS =
3q
2
_
3 + 2

3
_ _
2

3 1 q
_ _
q 2 +

3
_
2
8(q + 1)(q 2)
2
0.
The inequality is proved, and we conclude that k
max
= k
0
.
Mathematical Reflections 2 (2007) 5
2.9 [Pham Huu Duc] For all positive real numbers a, b and c,
1
a
2
+ bc
+
1
b
2
+ ca
+
1
c
2
+ ab

(a + b + c)
2
3(ab + bc + ca)
_
1
a
2
+ b
2
+
1
b
2
+ c
2
+
1
c
2
+ a
2
_
.
Solution. Because the inequality is homogeneous, we may assume that p = 1. Then q [0, 1] and by
the AM - GM and Schurs inequalities, we have
(1q
2
)
2
9
3r max
_
0,
14q
2
9
_
. After expanding, we can
rewrite the given inequality as
f(r) =486(9 q
2
)r
3
+ 27(q
6
+ 64q
4
35q
2
+ 24)r
2
+ 9(4q
2
1)(11q
4
4q
2
+ 2)r
+ q
2
(1 q
2
)
3
(2q
4
+ 8q
2
1) 0.
We have
f

(r) = 9(162(9 q
2
)r
2
+ 6(q
6
+ 64q
4
35q
2
+ 24)r + (4q
2
1)(11q
4
4q
2
+ 2))
f

(r) = 54(54(9 q
2
)r + q
6
+ 64q
4
35q
2
+ 24)
54(2(1 q
2
)
2
(9 q
2
) + q
6
+ 64q
4
35q
2
+ 24) = 162(q
6
+ 14q
4
+ q
2
+ 2) > 0.
Hence f

(r) is an increasing function.


Now, if 1 2q, then
f

(r) f

(0) = (4q
2
1)(11q
4
4q
2
+ 2) 0.
If 1 2q, then
f

(r) f

_
1 4q
2
27
_
= (1 4q
2
)(q
2
+ 2)(2q
4
+ 17q
2
+ 6) 0.
In any case, f(r) is an increasing function.
If 1 2q, then f(r) f(0) = q
2
(1 q
2
)
3
(2q
4
+ 8q
2
1) 0, and we are done. If 1 2q, using our
theorem, we have
f(r) f
_
(1 + q)
2
(1 2q)
27
_
=
1
81
q
2
(2 q)(q + 1)
2
(6q
3
+ 4q
2
7q + 4)(5q
2
2q + 2)
2
0.
The proof is complete. Equality holds if and only if a = b = c.
2.10 [Nguyen Anh Tuan] Let x, y, z be positive real numbers such that xy + yz + zx + xyz = 4. Prove
that
x + y + z
xy + yz + zx
1 +
1
48
((x y)
2
+ (y z)
2
+ (z x)
2
).
Solution. Since x, y, z > 0 and xy + yz + zx + xyz = 4, there exist a, b, c > 0 such that x =
2a
b+c
, y =
2b
c+a
, z =
2c
a+b
. The inequality becomes
P(a, b, c) =
(a + b + c)
2

cyc
(a
2
b
2
)
2
(a + b)
2
(b + c)
2
(c + a)
2

6

cyc
a(a + b)(a + c)

cyc
ab(a + b)
+ 12 0.
Because the inequality is homogeneous we can assume that p = 1. Then q [0, 1] and after some
computations, we can rewrite the inequality as
f(r) = 729r
3
+ 27(22q
2
1)r
2
+ 27(6q
4
4q
2
+ 1)r + (q
2
1)(13q
4
5q
2
+ 1) 0.
We have
f

(r) = 27(r(81r + 44q


2
2) + 6q
4
4q
2
+ 1).
Mathematical Reflections 2 (2007) 6
By Schurs inequality,
81r + 44q
2
2 3(1 4q
2
) + 44q
2
2 = 1 + 32q
2
> 0.
Hence f

(r) 0, and f(r) is an increasing function. Then by our theorem we have


f(r) f
_
(1 q)
2
(1 + 2q)
27
_
=
2
27
q
2
(q 1)(q + 2)
2
(4q
4
+ 14q
3
+ 15q
2
7q + 1) 0.
The inequality is proved. Equality holds if and only if x = y = z.
2.11 [Nguyen Anh Tuan] For all nonnegative real numbers a, b, c
_
(a
2
ab + b
2
)(b
2
bc + c
2
)+
_
(b
2
bc + c
2
)(c
2
ca + a
2
)+
_
(c
2
ca + a
2
)(a
2
ab + b
2
) a
2
+b
2
+c
2
.
Solution. After squaring both sides, we can rewrite the inequality as
2

cyc
(a
2
ab + b
2
)
_

cyc
_
a
2
ab + b
2
_

cyc
ab
__

cyc
a
2
_

cyc
a
2
b
2
.
By the AM - GM inequality,
_
a
2
ab + b
2

1
2
(a + b),
_
b
2
bc + c
2

1
2
(b + c),
_
c
2
ca + a
2

1
2
(c + a).
It suces to prove that
2

cyc
(a
2
ab + b
2
)
_

cyc
a
_

cyc
ab
__

cyc
a
2
_

cyc
a
2
b
2
.
Because this inequality is homogeneous, we can assume p = 1. Then q [0, 1] and the inequality is
equivalent to
2
_
72r
2
+ 3(1 10q
2
)r + q
2
(1 q
2
)
2
6r + q
2
(1 q
2
),
or
f(r) = 324r
2
12r(q
4
11q
2
+ 1) q
2
(4 q
2
)(1 q
2
)
2
0.
It is not dicult to verify that f(r) is a convex function, then using our theorem, we have
f(r) max
_
f(0), f
_
(1 q)
2
(1 + 2q)
27
__
.
Furthermore,
f(0) = q
2
(4 q
2
)(1 q
2
)
2
0,
f
_
(1 q)
2
(1 + 2q)
27
_
=
1
9
q
2
(q 1)
3
(q + 2)(9q
2
+ q + 2) 0.
Our proof is complete. Equality holds if and only if a = b = c or
a = t 0, b = c = 0, and their permutations.
Vo Quoc Ba Can: Can Tho University, Vietnam
Mathematical Reflections 2 (2007) 7

You might also like