Difference between revisions of "1983 IMO Problems/Problem 6"
(→Problem 6) |
(→Problem 6) |
||
(5 intermediate revisions by 3 users not shown) | |||
Line 1: | Line 1: | ||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
==Solution 1== | ==Solution 1== | ||
Line 36: | Line 29: | ||
which is the desired result. | which is the desired result. | ||
− | == | + | ==Solution 3== |
+ | Let <math>s</math> be the semiperimeter, <math>\frac{a+b+c}{2}</math>, of the triangle. Then, <math>a=s-\frac{-a+b+c}{2}</math>, <math>b=s-\frac{a-b+c}{2}</math>, and <math>c=s-\frac{a+b-c}{2}</math>. We let <math>x=\frac{-a+b+c}{2},</math> <math>y=\frac{a-b+c}{2}</math>, and <math>z=\frac{a+b-c}{2}.</math> (Note that <math>x,y,z</math> are all positive, since all sides must be shorter than the semiperimeter.) Then, we have <math>a=s-x</math>, <math>b=s-y</math>, and <math>c=s-z</math>. Note that <math>x+y+z=s</math>, so <cmath>a=y+z,b=x+z,c=x+y.</cmath> Plugging this into <cmath>a^2b(a-b)+b^2c(b-c)+c^2a(c-a)\geq0</cmath> and doing some expanding and cancellation, we get <cmath>2x^3z+2xy^3+2yz^3-2x^2yz-2xy^2z-2xyz^2\geq0.</cmath> The fact that each term on the left hand side has at least two variables multiplied motivates us to divide the inequality by <math>2xyz</math>, which we know is positive from earlier so we can maintain the sign of the inequality. This gives <cmath>\frac{x^2}{y}-x+\frac{y^2}{z}-y-z+\frac{z^2}{x}\geq0.</cmath> We move the negative terms to the right, giving <cmath>\frac{x^2}{y}+\frac{y^2}{z}+\frac{z^2}{x}\geq x+y+z.</cmath> We rewrite this as <cmath>\sum_{cyc}\frac{x^2}{y}\geq\sum_{cyc}rx+(1-r)y.</cmath> where <math>r</math> is any real number. (This works because if we evaulate the cyclic sum, then as long as the coefficients of <math>x</math> and <math>y</math> on the right sum to 1 the right side will be <math>x+y+z</math>. | ||
+ | |||
+ | ==Solution4== | ||
+ | "Solution from 111 problems in Algebra and Number Theory" |
Latest revision as of 06:40, 22 May 2024
Contents
Solution 1
By Ravi substitution, let ,
,
. Then, the triangle condition becomes
. After some manipulation, the inequality becomes:
.
By Cauchy, we have:
with equality if and only if
. So the inequality holds with equality if and only if x = y = z. Thus the original inequality has equality if and only if the triangle is equilateral.
Solution 2
Without loss of generality, let . By Muirhead or by AM-GM, we see that
.
If we can show that , we are done, since then
, and we can divide by
.
We first see that, , so
.
Factoring, this becomes . This is the same as:
.
Expanding and refactoring, this is equal to . (This step makes more sense going backwards.)
Expanding this out, we have
,
which is the desired result.
Solution 3
Let be the semiperimeter,
, of the triangle. Then,
,
, and
. We let
, and
(Note that
are all positive, since all sides must be shorter than the semiperimeter.) Then, we have
,
, and
. Note that
, so
Plugging this into
and doing some expanding and cancellation, we get
The fact that each term on the left hand side has at least two variables multiplied motivates us to divide the inequality by
, which we know is positive from earlier so we can maintain the sign of the inequality. This gives
We move the negative terms to the right, giving
We rewrite this as
where
is any real number. (This works because if we evaulate the cyclic sum, then as long as the coefficients of
and
on the right sum to 1 the right side will be
.
Solution4
"Solution from 111 problems in Algebra and Number Theory"