Maximum possible integer square root polynomial 1: easier (but still hard)

Find the maximum possible integral value of

25 x 2 103 x + 2014 \sqrt{25x^2-103x+2014}

Where x is an integer.

Assume:

I will post a solution to this with a neat generalized formula 16 hours after this post if no one else has.

You may use a calculator

You may use all and any means possible of solving this problem, whether be it wolfram alpha, Python coding, or Java. Although I encourage you to use only a pencil and a hand held calculator because that is how I did it.

the harder version of this is here


The answer is 3180.

This section requires Javascript.
You are seeing this because something didn't load right. We suggest you, (a) try refreshing the page, (b) enabling javascript if it is disabled on your browser and, finally, (c) loading the non-javascript version of this page . We're sorry about the hassle.

3 solutions

Patrick Corn
Oct 22, 2014

Let y = 25 x 2 103 x + 2014 y = \sqrt{25x^2-103x+2014} . Let a = 10 y , b = 50 x 103 a = 10y, b= 50x-103 . Then squaring , multiplying both sides by 100 100 , and completing the square gives ( a + b ) ( a b ) = 190791 (a+b)(a-b) = 190791 .

Note that a + b 7 a+b \equiv 7 mod 10 10 and a b 3 a-b \equiv 3 mod 10 10 . Note also that both factors should be positive because their sum is 20 y 20y , which is positive. There are six pairs of factors of 190791 = 3 2 17 29 43 190791 = 3^2\cdot 17 \cdot 29 \cdot 43 of this form: 63597 3 , 17 11223 , 4437 43 , 87 2193 , 1247 153 , 387 493. 63597 \cdot 3, 17 \cdot 11223, 4437 \cdot 43, 87 \cdot 2193, 1247 \cdot 153, 387 \cdot 493. This leads to the solutions ( x , y ) = (x,y) = ( 638 , 3180 ) , ( 110 , 562 ) , ( 46 , 224 ) , ( 19 , 114 ) , ( 13 , 70 ) , ( 1 , 44 ) . (638,3180), (-110,562), (46,224), (-19,114), (13,70), (1,44). So the answer is 3180 \fbox{3180} .

Nice alternative solution. Mine is a more case specific and formula based while yours is more intuitive and manipulative.

Trevor Arashiro - 6 years, 7 months ago

The graph of this function is an infinite parabola. So can you please explain to me as to how can it have a finite maximum value?

Sarthak Tanwani - 6 years, 7 months ago

Log in to reply

We're restricting to integer values of the domain and range.

Patrick Corn - 6 years, 7 months ago
Aneesh Kundu
Nov 3, 2014

Let k k be an integer such that 25 x 2 103 x + 2014 = k 2 25x^2-103x+2014=k^2 25 x 2 103 x + ( 2014 k 2 ) = 0 25x^2-103x+(2014-k^2)=0 Since x x is an integer satifying the above equation, the discrminant of the above quadratic is a perfect square .

Let p p be that perfect square . 10 3 2 4 ( 2014 k 2 ) ( 25 ) = p 2 103^2-4(2014-k^2)(25)=p^2 (expand it) 4 ( 25 ) ( 2014 ) 10 3 2 = ( 10 k ) 2 p 2 \Rightarrow 4(25)(2014)-103^2=(10k)^2-p^2 ( 10 k p ) ( 10 k + p ) = 190791 \Rightarrow (10k-p)(10k+p)=190791

We know that 190791 = 3 2 × 17 × 29 × 43 190791=3^2\times 17\times 29\times 43

Let α 1 \alpha_1 and α 2 \alpha_2 be factors of 190791 190791 , such that

( 10 k p ) = α 1 (10k-p)=\alpha_1 ( 10 k + p ) = α 2 (10k+p)=\alpha_2 α 1 α 2 = 190791 \alpha_1 \alpha_2=190791

Add the first 2 eqns

20 k = α 1 + α 2 20k=\alpha_1+\alpha_2

So we need to find 2 factors of 190791 such that there sum is maximum, by AM-GM we know that the minimal value would occur at symmetry(both the factors being equal), so we can think that at the extreme points the sum would maximize. We start by taking one of the factors to be unity(one) and the other to be the 190791, but we fail to get an integer solution. Second we take one factor to be 3 and the other to be 63579, in this case we get an an integer solution i.e, k = 3180 \boxed{k=3180}

Trevor Arashiro
Oct 20, 2014

let y = 25 x 2 103 x + 2014 y=\sqrt{25x^2-103x+2014}

Which represents the function a 2 x 2 + b x + c \sqrt{a^2x^2+bx+c}

Thus, because the first term is obviously a square number for an integer X, 103 x 2014 103x-2014 must be a difference of squares.

We can represent y 2 = 25 ( x + n ) 2 y^2=25(x+n)^2 implying that 25 ( x + n ) 2 25 x 2 = 103 x + 2014 25(x+n)^2-25x^2=-103x+2014

From here, a little manipulation yields x = 2014 25 n 2 103 + 50 n x=\dfrac{2014-25 n^2}{103+50 n}

Now, I don't know of any guarnteed way of finding an exact n without guess and check. Luckily, however, the max value of x occurs when the denominator is as small as possible, or 103 = 50 n 103=-50n

However, since n must be an integer for x to be an integer, we round to the nearest integer, thus n=-2. As our first guess, we can try n=-2, which, coincidentally, makes x an integer.

Thus plugging in n=-2, x=638. Plugging in x=638 we get y = 3180 \boxed{y=3180}

I will post a note for a much more generalized form of this.

Brilliant! I got to n=-103/50 but didn't understand what to do with that haha. Thanks!

Yash Akhauri - 6 years, 7 months ago

Log in to reply

Ya, it's cool that every number a,b,c,n has to be an integer or y is guarnteed irrational

Trevor Arashiro - 6 years, 7 months ago

At x=638, you will get maximum value of 3180. Please check your answer again! Btw I need my 165 points! :-|

Pranjal Jain - 6 years, 7 months ago

Log in to reply

Yeah, sorry about that, I didn't realize at the time that the lemma I used only holds for a,b,c being positive

Trevor Arashiro - 6 years, 7 months ago

Log in to reply

What are a,b,c??? Well... Can you share your lemma?

Pranjal Jain - 6 years, 7 months ago

Log in to reply

@Pranjal Jain I will post a note detailing it further , also the problem has been fixed so you should have been rewarded you points (assuming you put the answer as 3180.

Trevor Arashiro - 6 years, 7 months ago

Log in to reply

@Trevor Arashiro Yeah! I got points! Thanks btw!

Pranjal Jain - 6 years, 7 months ago

@Trevor Arashiro apologies, but shouldn't the maximum possible value of given polynomial by infinity since the coefficient of x^2 is greater than 0? Shouldn't there always be a number greater than 638 which gives an integral solution? Please elaborate...

Yash Akhauri - 6 years, 7 months ago

Log in to reply

@Yash Akhauri Well, you are inputting x as integers. So you might get infinite results but they need not be perfect squares to give y an integer!

Pranjal Jain - 6 years, 7 months ago

Log in to reply

@Pranjal Jain Could you provide me with a generalized way to solve such sums? Or this one specifically? Thanks"

Yash Akhauri - 6 years, 7 months ago

Log in to reply

@Yash Akhauri There is no exact general one, but I will post my solution right no.

Trevor Arashiro - 6 years, 7 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...