Solving equations (part 2)

Algebra Level 4

{ x + y + z = a , x 2 + y 2 + z 2 = b , x 3 + y 3 + z 3 = c , x 4 + y 4 + z 4 = d , \large \begin{cases} x+y+z=a,\\ x^2+y^2+z^2=b,\\ x^3+y^3+z^3=c, \\ x^4+y^4+z^4=d, \end{cases}

Given that a , b , c a, b, c and d d are integers and they are four consecutive terms in an arithmetic progression, where 0 a < b < c < d 0\le a<b<c<d . If the above system of equations has solutions (ie the system is consistent), how many such ( a , b , c , d ) (a,b,c,d) ?

2 3 0 1 4 \ge 4

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.

2 solutions

Yuriy Kazakov
Jul 2, 2019

I use Wolfram

Part 1 x + y + z = a , x 2 + y 2 + z 2 = b , x 3 + y 3 + z 3 = b + ( b a ) , f i n d x 4 + y 4 + z 4 x+y+z=a, x^2+y^2+z^2=b, x^3+y^3+z^3=b+(b-a), find \: \: x^4+y^4+z^4

Part 2 1 / 6 ( a 4 2 a 2 ( 3 b + 4 ) + 16 a b + 3 b 2 ) = b + 2 ( b a ) i n t e g e r s o l u t i o n 1/6 (a^4 - 2 a^2 (3 b + 4) + 16 a b + 3 b^2)=b+2(b-a) \: \: integer solution

a = 0 a n d b = 6 a = 0 \: \:and\: \: b = 6 \\ a = 3 a n d b = 5 a = 3 \: \:and\: \: b = 5 \\ a = 6 a n d b = 10 a = 6 \: \:and\: \: b = 10 \\ a = 6 a n d b = 36 a = 6 \: \:and\: \: b = 36 \\ a = 7 a n d b = 13 a = 7 \: \:and\: \: b = 13 \\

Chan Lye Lee
May 14, 2019

Partial solution:

It can be shown that the system is consistent if ( a , b , c , d ) = ( 0 , 6 , 12 , 18 ) , ( 3 , 5 , 7 , 9 ) , ( 6 , 10 , 14 , 18 ) , ( 7 , 13 , 19 , 25 ) (a,b,c,d)=(0,6,12,18), (3,5,7,9), (6,10,14,18), (7, 13, 19, 25) .

How can the numbers 7,13,20,27 be in A.P.?

Log in to reply

It was a typo. It should be ( a , b , c , d ) = ( 7 , 13 , 19 , 25 ) (a,b,c,d)=(7, 13, 19, 25) .

Chan Lye Lee - 2 years ago

I don't get it, b+d-2c=0, since b,c,d are in AP. That means sum(x⁴-2x³+x²)=0 => sum(x²(x-1)²)=0. Since x,y,z are all reals, and sum of square of real numbers is 0, it means the numbers itself ard 0, that means x,y,z €{0,1}. Now we can just check cases and see that none satisfy.

Kushal Dey - 4 months, 2 weeks ago

Log in to reply

you wrote sum(x⁴-2x³+x²)=0. I think that means b + d 2 c = x 4 2 x 3 + x 2 + y 4 2 y 3 + y 2 + z 4 2 z 3 + z 2 = 0 b+d-2c=x^4-2x^3+x^2 + y^4-2y^3+y^2 + z^4-2z^3+z^2=0 x 2 ( x 1 ) 2 + y 2 ( y 1 ) 2 + z 2 ( z 1 ) 2 = 0 \Rightarrow x^2(x-1)^2 + y^2(y-1)^2 + z^2(z-1)^2=0 That seems true to me, but why do x, y and z have to be real?

K T - 4 months, 2 weeks ago

Log in to reply

@K T Yes, I exactly meant that and x,y,z in my opinion should be real because we were only asked to confirm the consistency of the system of equations and nothing else was mentioned in the question, in that case we usually don't assume complex values.

Kushal Dey - 4 months, 1 week ago

( a , b , c , d ) = ( 6 , 36 , 66 , 96 ) (a,b,c,d)=(6,36,66,96) is another valid solution, so there are (at least?) 5 5 such consistent systems (not that this changes the answer).

My workings were pretty messy - lots of Vieta, a bit too much reliance on Wolfram|Alpha...would you be able to post a full solution?

Chris Lewis - 2 years ago

Log in to reply

From the given equations, xy+yz+zx= a 2 b 2 \dfrac{a^2-b}{2} , xyz= 1 6 \dfrac{1}{6} (2c-3ab+ a 3 a^3 ). Therefore x, y, z are the roots of the equation X 3 X^3 -a X 2 X^2 + 1 2 \dfrac{1}{2} ( a 2 a^2 -b) X- 1 6 \dfrac{1}{6} ( a 3 a^3 -3ab+2c)=0. Therefore d= x 4 x^4 + y 4 y^4 + z 4 z^4 =a( x 3 x^3 + y 3 y^3 + z 3 z^3 )- 1 2 \dfrac{1}{2} ( a 2 a^2 -b) ( x 2 x^2 + y 2 y^2 + z 2 z^2 )+ 1 6 \dfrac{1}{6} ( a 3 a^3 -3ab+2c)(x+y+z). Substituting the values, we get the equation I have written.

Log in to reply

Yep, I agree with your equation, but the question is about the number of solutions it has in integers with a , b , c , d a,b,c,d in A.P. Of course, we can just substitute in the five integer AP solutions that have been listed so far and show that they solve the equation, and that's enough to answer the question; but I'd be interested in showing that these are the only solutions.

Chris Lewis - 2 years ago

Log in to reply

@Chris Lewis Piggybacking on both the author's work and your work, the quadratic discriminant (in w w ) of the equation a 4 6 a 3 6 a 2 w + 11 a 2 + 22 a w 6 a + 3 w 2 18 w = 0 a^4 - 6 a^3 - 6 a^2 w + 11 a^2 + 22 a w - 6 a + 3 w^2 - 18 w = 0 must be a perfect square. This essentially tells us that we want to solve the Diophantine equation, 6 a 4 48 a 3 + 142 a 2 180 a + 81 = E 2 . 6 a^4 - 48 a^3 + 142 a^2 - 180 a + 81 = E^2 .

I checked WolframAlpha for a < 1 0 10 a < 10^{10} and it only displays the 5 solutions already presented:

( a , b , c , d ) = ( 0 , 6 , 12 , 18 ) , ( 3 , 5 , 7 , 9 ) , ( 6 , 10 , 14 , 18 ) , ( 6 , 36 , 66 , 96 ) , ( 7 , 13 , 19 , 25 ) . (a,b,c,d) = (0,6,12,18), (3,5,7,9), (6,10,14,18), (6,36,66,96), (7, 13, 19, 25).

I doubt there's a "clean" solution to this problem.

Pi Han Goh - 2 years ago

Log in to reply

@Pi Han Goh Yeah... I'm a bit out of my depth with these, but it's very enticing (especially since all the solutions have a < 10 a<10 , which suggests it's manageable).

Of course, the next question is, what about generalising? How many consistent systems of the form

i = 1 n x i j = a j \sum_{i=1}^n x_i^j=a_j where j = 1 , , n + 1 j=1,\ldots,n+1 and the a j a_j are distinct non-negative integers in arithmetic progression are there? (It looks like the answer is 5 5 when n = 3 n=3 )

Chris Lewis - 2 years ago

I got the point. Thanks. May be there are more than five solutions. I'm checking it now.

From the given equations, it can be shown that a 4 6 a 2 b + 8 a c + 3 b 2 6 d = 0. a^4 - 6a^2 b + 8ac + 3b^2 - 6d = 0. Then using the arithmetic progression condition, this reduces to a 4 6 a 2 b 8 a 2 + 16 a b + 3 b 2 + 12 a 18 b = 0. a^4 - 6a^2 b - 8a^2 + 16ab + 3b^2 + 12a - 18b = 0.

Jon Haussmann - 2 years ago

Log in to reply

Any ideas how to find the integer solutions from this? If you treat it as a quadratic in b b , you get the same discriminant as @Pi Han Goh derived above (where the two parameters considered were the initial term and the common difference of the arithmetic progression). But there might be something in this formulation (or another one) that comes out more clearly.

Chris Lewis - 2 years ago

Log in to reply

@Chris Lewis , @Pi Han Goh , I am just curious why I cant view mention[54224:Pi Han Goh]'s comments?

Chan Lye Lee - 2 years ago

Log in to reply

@Chan Lye Lee The comment I was referring to is partway through the replies to my comment about the fifth solution

Looking again, it starts with an "at" symbol - does that make it private? If you can't see a reply to a comment starting with "Piggybacking" let me know and I can post it here (or else try to summarise findings in a solution if easier).

Chris Lewis - 2 years ago

Log in to reply

@Chris Lewis @Chris Lewis I got it. Thanks.

Chan Lye Lee - 2 years ago

Nope, no idea where to go from here.

Jon Haussmann - 2 years ago

I have no proof, only an idea to make the search a bit more efficient and avoid huge numbers evaluating the discriminant. By using modular arithmetic modulo a prime p, you can exclude all values of a a for which the discriminant D = 6 a 4 48 a 3 + 142 a 2 180 a + 81 ( m o d p ) D=6a^4 -48a^3+142a^2-180a+81 \pmod p does not evaluate to a quadratic residue. Combining the primes upto 150, I tested all integer values from 0 upto 10^6, and the only ones among these for which D is a perfect square are [0, 1, 2, 3, 6, 7].

K T - 1 year, 3 months ago

Here's a solution with a bit more detail:

We have a 3 = x 3 + y 3 + z 3 + 3 ( x y 2 + y z 2 + z x 2 + x 2 y + y 2 z + z 2 x ) + 6 x y z a^3=x^3+y^3+z^3+3(xy^2+yz^2+zx^2+x^2y+y^2z+z^2x) +6xyz , so a 3 3 a b + 2 c 6 x y z = 0 (I) a^3-3ab+2c-6xyz=0 \text{ (I)}

Similarly we have a 4 = x 4 + y 4 + z 4 + 4 ( x y 3 + y z 3 + z x 3 + x 3 y + y 3 z + z 3 x ) + 6 ( x 2 y 2 + y 2 z 2 + z 2 x 2 ) + 12 a x y z a^4 =x^4+y^4+z^4+4(xy^3+yz^3+zx^3+x^3y+y^3z+z^3x)+6(x^2y^2+y^2z^2+z^2x^2) + 12axyz

= d + 4 ( a c d ) + 3 ( b 2 d ) + 12 a x y z =d+4(ac-d)+3(b^2-d) + 12axyz , so

a 4 4 a c 3 b 2 + 6 d 12 a x y z = 0 (II) a^4-4ac-3b^2+6d-12axyz=0 \text{ (II)}

Eliminate x y z xyz by multiplying ( I ) (I) by 2 a 2a and then subtract ( I I ) (II) from it:

a 4 6 a 2 b + 8 a c + 3 b 2 6 d = 0 a^4-6a^2b+8ac+3b^2-6d=0

Using the arithmetic progression, now substitute c = 2 b a c=2b-a and d = 3 b 2 a d=3b-2a and simplify to

a 4 6 a 2 b 8 a 2 + 16 a b + 3 b 2 + 12 a 18 b = 0 a^4 - 6a^2b - 8a^2 + 16ab + 3b^2 + 12a - 18b = 0

We can view this as a quadratic in b:

( 3 ) b 2 + ( 6 a 2 + 16 a 18 ) b + ( a 4 8 a 2 + 12 a ) = 0 (3)b^2  +( -6a^2+16a -18)b +(a^4 -8a^2+12a) = 0

with solutions b = 3 a 2 8 a + 9 ± ( 3 a 2 + 8 a 9 ) 2 3 ( a 4 8 a 2 + 12 a ) 3 b=\frac{3a^2-8a +9 \pm \sqrt{ (-3a^2+8a -9)^2 - 3(a^4 -8a^2+12a )}} {3}

in which the discriminant simplifies to

D = 6 a 4 48 a 3 + 142 a 2 180 a + 81 D = 6a^4 -48a^3+142a^2-180a+81 D needs to be a perfect square, so we can check different values of a, and find that D is a square for a { 0 , 1 , 2 , 3 , 6 , 7 } a \in \{0,1,2,3,6,7\} . This is not exhaustive, but by checking quadratic residues mod p for a prime p, I could exclude many candidates at a time. Doing so for all primes p 150 p \leq 150 , I sieved out all other candidate values a 1 , 000 , 000 a\leq 1,000,000 .

Checking the found values for a a , we find 5 solutions for ( a , b ) (a,b) :

  • Case a = 0 a=0 : b = 9 ± 9 3 b= \frac{9 \pm 9}{3} , of which b = 6 b=6 is a possible solution satisfying b > a b>a .
  • Case a = 1 a=1 : b = 4 ± 1 3 b= \frac{4 \pm 1}{3} , of which b = 1 b=1 is the only integer solution.
  • Case a = 2 a=2 : D = 1 , b = 12 16 + 9 ± 1 3 D=1, b=\frac{12-16+9 \pm 1}{3} , with integer solution b = 2 b=2 . Because of the condition b > a b \gt a , this is no valid solution.
  • Case a = 3 a=3 : D = 3 2 , b = 27 24 + 9 ± 3 3 D=3^2, b=\frac{27-24+9  \pm 3}{3} , with solutions b = 3 b=3 or b = 5 b=5 , of which only b = 5 b=5 satisfies b > a b>a .
  • Case a = 6 a=6 : D = 3 9 2 , b = 108 48 + 9 ± 39 3 D=39^2, b=\frac{108-48+9 \pm 39}{3} , with integer solutions b = 10 b=10 and b = 36 b=36 .
  • Case a = 7 a=7 : D = 6 1 2 , b = 147 56 + 9 ± 61 3 D=61^2, b=\frac{147-56+9 \pm 61}{3} , with integer solution b = 13 b=13 .
  • c c and d d follow directly from a a and b b .
  • We have not proved yet that to each a quadruple ( a , b , c , d ) (a,b,c,d) indeed one or more triples ( x , y , z ) (x, y, z) exists, but Wolfram alpha could find values for each of the above cases (even providing an analytic solution for the case a = 7 a=7 ).

Summarized we found solutions ( a , b , c , d ) { ( 0 , 6 , 12 , 18 ) , ( 3 , 5 , 7 , 9 ) , ( 6 , 10 , 14 , 18 ) , ( 6 , 36 , 66 , 96 ) , ( 7 , 13 , 19 , 25 ) } (a,b,c,d) \in \{(0,6,12,18),(3,5,7,9),(6,10,14,18),(6,36,66,96),(7,13,19,25)\}

K T - 1 year, 3 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...