a1/a2 + a2/a3 + a3/a4 + ... + an/a1 >= a1/P + a2/P + ... an/P >= n/P >= n. (qed)
( with P = a1.a2.a3... an. )
2006-07-17 01:22:55
·
answer #1
·
answered by gjmb1960 7
·
0⤊
1⤋
Starting with the first equation...
a/b + b/a = a^2/ab + b^/ab = (a^2 + b^2)/ab
Let us say that a >= b for the sake of convenience (it works for b > a as well). Thus a = b + n, where integer n >= 0.
a/b + b/a = [(b + n)^2 + b^2]/[(b + n)b]
= (2b^2 + 2bn + n^2)/(b^2 + bn)
= 2(b^2 + bn)/(b^2 + bn) + n^2/(b^2 + bn)
= 2 + n^2/(b^2 + bn)
Since the fraction must be positive or zero, then the proof holds that a/b + b/a >= 2
The other equations are solved similarly, but I'll let you work on those. It's probably a matter of applying a recursive proof.
2006-07-17 00:51:34
·
answer #2
·
answered by Ѕємι~Мαđ ŠçїєŋŧιѕТ 6
·
0⤊
0⤋
Yes. The numbers only need to be positive, not integers. This can be proved by induction on the number of terms n in the sum. Obviously if there's only one term:
S = a/a = 1 >= 1
So assume the inequality holds for such sums involving n-1 terms and consider the sum:
S = a1/a2 + a2/a3 + ... + a_(n-1)/an + an/a1
Without loss of generality we can assume that a1 >= ai for all 1 <= i <= n. (In other words if originally ai = Max{a1,...an} but i != 1, shift the first i-1 terms to the end of the sum and relabel all the a's cyclically, ai-->a1, a_(i+1)--> a2, etc..)
Then a1/a2 >= 1 and an/a1 <= 1 so that:
(a1/a2 - 1)*(1 - an/a1) >= 0
-->
a1/a2 - 1 - an/a2 + an/a1 >= 0
-->
a1/a2 + an/a1 >= 1 + an/a2
The left side of the above inequality is the first and last terms in S, so we can replace them by the right side (1+an/a2):
S >= a2/a3 +...+ a_(n-1)/an + (1 + an/a2)
= (a2/a3 + a3/a4 + ... + an/a2) + 1
>= n-1 + 1 >= n
using the inductive hypothesis.
2006-07-17 08:30:01
·
answer #3
·
answered by shimrod 4
·
0⤊
0⤋
a,b,c,d,…. all are integers so a >= 1, b >= 1, c >= 1, d >=1, ….. and so on.
so ab â 0; so ab/ab = 1;
equation is ( a/b + b/a ) >= 2
multiply the LHS by (ab/ab) that is one
LHS = ab/ab ( a/b + b/a)
= ab( a/b + b/a) /ab
= (a2 + b2) / ab
= (a2 + b2 - 2ab + 2ab) / ab
= ( (a-b) 2 + 2ab) / ab
= (a-b)2 //ab + 2
LHS= ( a – b )2 / ab + 2
( a – b ) 2 >= 0 and ab > 1
so ( a – b ) 2 / ab >= 0
so ( a – b ) 2 / ab + 2 >= 2
so LHS >= 2
like this u can multiply other equations by abc/abc, abcd/abcd, …. and so on
2006-07-17 01:24:42
·
answer #4
·
answered by Nasmi 2
·
0⤊
0⤋
Why don't you do the simple thing and apply a value to the unknowns?
a=4
b=2
Using " / ", you are indicating division, aren't you?
4/2+2/4...............=2 ??????????????????HUH?
Am I missing something? I have both oars in the water, I think.
2006-07-17 00:51:36
·
answer #5
·
answered by ed 7
·
0⤊
0⤋
For the first one:
(a-b)(a-b)>=0 is true Then
a*a+b*b-2a*b>=0
a*a+b*b>=2*a*b
(a*a+b*b)/(a*b)>=2
By write this using two fractions, you will get
a/b+b/a>=2
2006-07-17 00:51:31
·
answer #6
·
answered by iyiogrenci 6
·
0⤊
0⤋
see multiply the l.h.s wit the l.c.m n u get the proof...
2006-07-17 00:42:31
·
answer #7
·
answered by partha s 2
·
0⤊
0⤋
umm do your own homework!
2006-07-17 00:39:53
·
answer #8
·
answered by Kittie_Nash 5
·
0⤊
0⤋
say what?????????
2006-07-17 00:39:56
·
answer #9
·
answered by Anonymous
·
0⤊
0⤋