Вы находитесь на странице: 1из 1

Mathematical Excalibur, Vol. 11, No. 1, Feb. 06 - Mar.

06 Page 2

Example 3. (1998 IMO Shortlisted (1) [(9,0,0)] [(7,1,1)] [(6,0,0)] c < d. Let r = (r1,,rn) be defined by ri
Problem) For any x, y, z > 0 with xyz = = pi except rj = b + c and rk = b c. By
1, prove that (2) [(7,5,0)] [(5,5,2)] the definition of c, either rj = qj or
3 3 3
rk=qk. Also, by the definitions of b, c,
x y z 3 (3) 2[(7,5,0)] 2[(6,5,1)] 2[(5,4,0)] d, we get p r, p r and r q. Now
+ + .
(1+ y)(1+ z) (1+ z)(1+ x) (1+ x)(1+ y) 4
(4) [(7,5,0)] + [(5,2,2)] 2[(6,7/2,1)]
x (x
pj r
Solution. Multiplying by the common n!([ p] [r]) = x pk
( j ) (k ) xj( j ) xrk(k ) )
denominator and expanding both sides, 2[(11/2,7/2,3/2)] 2[(4,2,0)] S n

the desired inequality is


(5) [(5,5,5)] [(2,2,2)], = x (u b+d
v b d u b + c v b c ),
4(x4+y4+z4+x3+y3+z3) S n

where (1) and (3) are by Muirheads


3(1+x+y+z+xy+yz+zx+xyz). inequality and the remark, (2) is by where x is the product of x p ( i ) for i

Muirheads inequality, (4) is by the fact, i j, k and u = x(j) , v = x(k). For each
This is equivalent to 4[(4,0,0)] + Muirheads inequality and the remark and
4[(3,0,0)] [(0,0,0)] + 3[(1,0,0)] + permutation , there is a permutation
(5) is by the remark.
3[(1,1,0)] + [(1,1,1)]. such that (i) = (i) for i j, k and (j)
Considering the sum of the leftmost parts = (k), (k) = (j). In the above sum, if
For this, we apply Muirheads of these inequalities is greater than or we pair the terms for and , then x =
inequality and the trick as follow: equal to the sum of the rightmost parts of x and combining the parenthetical
these inequalities, we get the desired factors for the and terms, we have
[(4,0,0)] [(4/3,4/3,4/3)] = [(0,0,0)], inequalities.
3[(4,0,0)] 3[(2,1,1)] = 3[(1,0,0)], (ub+d vbd ub+c vbc)+(vb+d ubd vb+c ubc)
3[(3,0,0)] 3[(4/3,4/3,1/3)] = 3[(1,1,0)] = ubd vbd (ud+c vd+c) (udc vdc) 0.
Alternate Solution. Since
and [(3,0,0)] [(1,1,1)] . So the above sum is nonnegative. Then
x x
5 2
x x
5 2 [p] [r]. Equality holds if and only if u
Adding these, we get the desired
= v for all pairs of and , which yields
inequality. x5 + y 2 + z 2 x3 ( x 2 + y 2 + z 2 )
x1= x2 = = xn. Finally we recall r has
( x 3 1) 2 ( y 2 + z 2 ) at least one more coordinate in
Remark. For the following example, = 0, agreement with q than p. So repeating
we will modify the trick above. In case x( x + y 2 + z 2 )( x 5 + y 2 + z 2 )
2
this process finitely many times, we
xyz 1, we have
we have will eventually get the case r = q. Then
[(p1, p2, p3)] [(p1r, p2r, p3r)] we are done.
x5 x 2 y5 y 2 z5 z2
+ +
for every r 0. Also, we will use the x5 + y 2 + z 2 y5 + z 2 + x2 z 5 + x2 + y 2 Next, for the advanced readers, we
following will outline a longer proof, which tells
x5 x2 y5 y2 z5 z 2 more of the story. It is consisted of two
+ 3 2 2 2 + 3 2 2 2
Fact. For p, q n, we have x (x + y + z ) y ( y + z + x ) z (z + x + y )
3 2 2 2
steps. The first step is to observe that if
[ p] + [q] p + q c1, c2, , ck 0 with sum equals 1 and
. 1 1 1 1 v1, v2, , vk n, then
2 2
x2 + y2 + z 2
(x2 + y2 + z 2 )
x y z k
k
This is because by the AM-GM c [v i ] c i v i .
i
i =1
inequality, 1 i =1
( x 2 + y 2 + z 2 yz zx xy )
x p1
Lx pn
+xq1
Lx qn x2 + y2 + z2 This follows by using the weighted
(1 ) (n) (1 ) (n)
AM-GM inequality instead in the proof
2 ( x y ) 2 + ( y z ) 2 + ( z x) 2 of the fact above. (For the statement of
= 0.
x( p(11+) q1 ) / 2 L x( p( nn+) qn ) / 2 . 2( x 2 + y 2 + z 2 ) the weighted AM-GM inequality, see
Mathematical Excalibur, vol. 5, no. 4,
Summing over Sn and dividing by n!, p. 2, remark in column 1).
we get the inequality.
Proofs of Muirheads Inequality The second step is the difficult step of
Example 4. (2005 IMO) For any x, y, z showing p q implies there exist
> 0 with xyz 1, prove that Kin Yin Li nonnegative numbers c1, c2, , cn! with
sum equals 1 such that
x5 x2 y5 y 2 z5 z 2 Let p q and p q. From i = 1 to n, the
+ + 0. n!
x5 + y2 + z 2 y5 + z 2 + x2 z5 + x2 + y2 first nonzero pi qi is positive by q = ci Pi ,
i =1
Solution. Multiplying by the common condition 2 of majorization. Then there is
denominator and expanding both sides, a negative pi qi later by condition 3. It where P1, P2, , Pn! n whose
the desired inequality is equivalent to follows that there are j < k such that pj > qj, coordinates are the n! permutations of
[(9,0,0)]+4[(7,5,0)]+[(5,2,2)]+[(5,5,5)] pk < qk and pi = qi for any possible i the coordinates of p. Muirheads
[(6,0,0)] + [(5,5,2)] + 2[(5,4,0)] + between j, k. inequality follows immediately by
2[(4,2,0)] + [(2,2,2)]. applying the first step and observing
Let b = (pj+pk)/2, d = (pjpk)/2 so that that [Pi]=[p] for i=1,2,, n!.
[bd,b+d] = [pk, pj] [qk, qj]. Let c be
To prove this, we note that (continued on page 4)
the maximum of |qjb| and |qkb|, then 0

Вам также может понравиться