math problem for the sake of it.

sao123

Lifer
May 27, 2002
12,653
205
106
Go Here(pdf file)

... and read section 2, i believe it will shed some insight on how to solve this problem. Use the case Q < 0.

Pretty heavy stuff there.

Lemme know if this proves useful.
 

uart

Member
May 26, 2000
174
0
0
Well Dr Pizza, you'll be pleased to know that working independantly (didn't even look at your working) I arrived at the same final answer as yourself, so I think you are correct. In particular the bottom line that I got was :

For each value of m in the range 4 * (600)^2 < m < 5 * (600)^2 there is a suitable unique value of n provided that 5 * (600)^2 - m is not a perfect square.

This gave me a total of (600)^2 - 1 - ( 600 - 1) = 359400 possibilities, exactly the number you arrived at.

BTW, n = 2 * 600 * ( 4 * 600^2 - m ), so each of these values of n is negative, not that the question actually asked that though.

Hmmm, I suppose it's possible that the examiners messed up on that question and didn't get the correct answer themselves.
 

uart

Member
May 26, 2000
174
0
0
BTW, It is interesting to note that all of the suggested possible answers on the paper are approx (500)^2 whereas the answer that both youself and I obtained is approx (600)^2[b/].

Hey I still think that they (the examiners) goofed, hopefully we'll find out eventually.
 

DrPizza

Administrator Elite Member Goat Whisperer
Mar 5, 2001
49,601
167
111
www.slatebrookfarm.com
Originally posted by: uart
Well Dr Pizza, you'll be pleased to know that working independantly (didn't even look at your working) I arrived at the same final answer as yourself, so I think you are correct. In particular the bottom line that I got was :

For each value of m in the range 4 * (600)^2 < m < 5 * (600)^2 there is a suitable unique value of n provided that 5 * (600)^2 - m is not a perfect square.

This gave me a total of (600)^2 - 1 - ( 600 - 1) = 359400 possibilities, exactly the number you arrived at.

BTW, n = 2 * 600 * ( 4 * 600^2 - m ), so each of these values of n is negative, not that the question actually asked that though.

Hmmm, I suppose it's possible that the examiners messed up on that question and didn't get the correct answer themselves.

That does make me feel better. I wasted 2 hours trying to find my mistake.
 

Sahakiel

Golden Member
Oct 19, 2001
1,746
0
86
Originally posted by: DrPizza
Hmm... no one else seems interested in figuring out what I did wrong. Oh well.

We're engineers. We don't do number theory. We take the results and do something useful with it...
 

uart

Member
May 26, 2000
174
0
0
Ok Dr Pizza, I'll outline how I approached the problem for the sake of comparison.

Step_1.

I denoted the three positive real roots as a, b and c and took a to be the integer root.

Step_2.

I did a polynimial division of,
P(x) = X^3 - 2400x^2 + mx + n
by (x-a) to get a quotient polynomial of,
Q(x) = x^2 + (a - 2400) x + ( m - 2400a + a^2)

Step_3.

Realizing that "b" and "c" are the roots of Q(x) I deduced that,
-(b+c) = ( a - 2400) .

But since a=b+c is given as data then -a = (a - 2400) or a=1200. So the integer root is now fully determined.

Step_4.

Subsituting the now known value for "a" back into the quadratic Q(x) easily gets the other two roots in the form of,
b,c = 600 +/- sqrt ( 5 * 600^2 - m )

Step_5.

Since "b" and "c" must be both positive and real it is neccessary that the argument of the above sqrt be both non-negative and not greater than 600^2. That is,
5 * 600^2 >= m >= 4 * 600 ^2 ,
giving at most 1 + 600^2 possible values for m.

Step_6.

Since the question stipulates that "a" is the only integer root then m is further restricted in that the argument of the sqrt in Step_4 can not be a perfect square. That is,
5 * 600^2 - m cannot be a perfect square. But since 5 * 600^2 - m can only be in the range of [0 .. 600^2] due to the restrictions of Step_5 then this gives precisely 601 possible perfect squares.

Step_7.

The net results is that there are precisely 1 + 600^2 - 601 (or 600^2 - 600) possible values of m

Step_8.

By expansion of the factored form of the original polynomial it is clear that,
a * b * c = -n.

That is, -n = 1200 * ( 600^2 - ( 5 * 600^2 - m ) ),

or simplifying, -n = 1200 * ( m - 4 * 600^2).

So clearly each value of m in the above specified range corresponds to a unique value of n and so there are also precisely 600^2 - 600 possible values for n. #



I hope that helps.
 

DrPizza

Administrator Elite Member Goat Whisperer
Mar 5, 2001
49,601
167
111
www.slatebrookfarm.com
So, you and I both got the same answer, which wasn't one of the multiple choices. Maybe the OP of that problem didn't type the choices exactly. Oh well.
 

sao123

Lifer
May 27, 2002
12,653
205
106
I did a polynimial division of,
P(x) = X^3 - 2400x^2 + mx + n
by (x-a) to get a quotient polynomial of,
Q(x) = x^2 + (a - 2400) x^2 + ( m - 2400a + a^2)

I performed this polynomial division myself, which results in a remainder, which doesnt appear you dealt with.
How did you ensure you actually have the integer root?
I believe the remainder when set to zero (to prove a is the integer) will provide you with a restriction on the value of N.

 

sao123

Lifer
May 27, 2002
12,653
205
106
furthermore doing long poly division by
x - 1200 which we have verified as a root leaves an remainder N - M + 1200 ^ 2 which in turn must be zero for 1200 to be the integer root. Leaving the equation N = M - 1200 ^ 2 which we must satisfy also.
 

uart

Member
May 26, 2000
174
0
0
Originally posted by: sao123

I performed this polynomial division myself, which results in a remainder, which doesnt appear you dealt with.
Yes I did, the remainder is always zero when you divide by a root factor.

In general you have, P(x) = Q(x) (x-a) + R (where Q is the quotient and R the remainder). It follows (by substituting x=a in the above) that R = P(a). This is a well know result called the Polynomial remainder theorem. It is true in general (whether or not "a" is a root), but naturally for the case when "a" is a root we simply have R = P(a)[/b] which is zero by definition.


How did you ensure you actually have the integer root?
We were told that one of the three roots was integer, I simply choose to name that root as "a". There is no loss of generality involved here. Eventually I was able to prove that a was in fact 1200, no doubt an integer.

furthermore doing long poly division by
x - 1200 which we have verified as a root leaves an remainder N - M + 1200 ^ 2 which in turn must be zero for 1200 to be the integer root. Leaving the equation N = M - 1200 ^ 2 which we must satisfy also.
No, you made an error in your calcuations. If you re-do that calc you will suely find that the remainder is,
R = n + 1200m - 1200^3 .

You may do this the hard way if you wish, you'll get the same answer (eventually), but remember that by the above "remainder theorem" that this remainder is none other than P(1200). That is,the original polynomial evaluated at x=1200.

Oh and btw, just take a look at step_8 in my solution and you will see precisely that equation, and yes it was set to equal zero as required.

 

DrPizza

Administrator Elite Member Goat Whisperer
Mar 5, 2001
49,601
167
111
www.slatebrookfarm.com
Originally posted by: sao123
furthermore doing long poly division by
x - 1200 which we have verified as a root leaves an remainder N - M + 1200 ^ 2 which in turn must be zero for 1200 to be the integer root. Leaving the equation N = M - 1200 ^ 2 which we must satisfy also.

n = 2 * 600 * ( 4 * 600^2 - m ) in UART's post

That condition was satisfied above and elsewhere...
 

sao123

Lifer
May 27, 2002
12,653
205
106
I found my math mistake, n + 1200m - 1200^3 is correct.

Bank error in your favor, Collect $200.
 

Brucmack

Junior Member
Oct 4, 2002
21
0
0
I played with this too, but couldn't think of anything that would limit the answer more.

However, it is interesting to note that the correct answer is 250,500. AMC Competition Results (question 23 on 12B)

It is also possible that the OP got the question wrong. Or that all of us are missing something...

 

uart

Member
May 26, 2000
174
0
0
Thanks for the info Bruce, that helps alot. Given that answer I worked backwards and figured that the polynomial in the original question was probably, x^3 - 2004 x*2 + mx +n and not, x^3 - 2400 x*2 + mx +n as given to us.

I've posted this in the original topic (OT) in the hope that the guy who posted the original question can confirm that he goofed in the transcription.
 

Brucmack

Junior Member
Oct 4, 2002
21
0
0
You are correct. I wrote someone about it and they gave 2004 as the correct value there. Makes sense too, since they normally use the year in some of the questions.