Suppose f ( x ) is a fixed polynomial with integer coefficients, such that for some integer n and m , f ( n ) = 0 and f ( m ) = 2 . Under these assumptions, what is the largest possible number of integers x such that f ( x ) = 6 ?
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.
Yes, this is basically correct. I've written down a similar solution (unpublished due to having too many solutions this week already :) ). Note that taking g ′ ( x ) = − f ( x ) + 6 is even better as you get rid of some minuses.
For the alternative proof that four factors won't work, it's enough to observe that when ± 1 , ± 2 are all factors then we actually have g ( x ) = − ( x 2 − 1 ) ( x 2 − 4 ) and the value where the polynomial takes the value − 4 is 0 . But now all of the integers less than 3 have been taken and the value of this polynomial is at most − 4 0 for ∣ x ∣ ≥ 3 so it cannot equal − 6 at any other integer.
Correction: that g ( 0 ) = − 4 caps the number of integer roots of g ( x ) at 6 .
Since the polynomial f ( x ) takes the value 0 for some integer, we can assume (by replacing f ( x ) by f ( x − u ) for some integer u if necessary) that f ( 0 ) = 0 , and hence f ( x ) = x g ( x ) for some polynomial g ( x ) ∈ Z [ x ] . There exists an integer v such that 2 = f ( v ) = v g ( v ) , and hence v ∣ 2 . By replacing f ( x ) by f ( − x ) if necessary, we can assume that v = 1 or 2 .
If v = 1 then f ( x ) = x [ ( x − 1 ) h ( x ) + 2 ] for some h ( x ) ∈ Z [ x ] . If w ∈ Z is such that f ( w ) = 6 , then w must divide 6 , and w cannot be 1 . For the different possible values of w , there are matching required values of h ( w ) for f ( w ) to be equal to 6 : w 6 3 2 − 1 − 2 − 3 − 6 h ( w ) − 5 1 0 1 4 3 5 1 7 3 and so w cannot be 6 , − 2 or − 6 . It is not possible to find a polynomial h ( x ) ∈ Z [ x ] such that h ( 3 ) = 0 , h ( 2 ) = 0 , h ( − 1 ) = 4 and h ( − 3 ) = 1 (if h ( 3 ) = 0 then 6 must divide h ( − 3 ) ), but h ( x ) = 3 − x satisfies three out of four of these conditions (at x = 3 , 2 , − 1 ).
If v = 2 then f ( x ) = x [ ( x − 2 ) h ( x ) + 1 ] for some h ( x ) ∈ Z [ x ] . Again, if w ∈ Z is such that f ( w ) = 6 , then w ∣ 6 and w = 2 , and we have the following table of required values of h ( w ) : w 6 3 2 − 1 − 2 − 3 − 6 h ( w ) 0 1 − 5 3 7 1 5 3 4 1 and so w cannot be − 1 , − 3 or − 6 . It is not possible to find a polynomial h ( x ) ∈ Z [ x ] such that h ( 6 ) = 0 , h ( 3 ) = 1 , h ( 2 ) = − 5 and h ( − 2 ) = 1 (if h ( 6 ) = 0 then 8 divides h ( − 2 ) ).
Thus we deduce that the most number of integers w such that f ( w ) = 6 is 3 , with f ( x ) = x [ ( x − 1 ) ( 3 − x ) + 2 ] = x ( − x 2 + 4 x − 1 ) achieving the value 6 for x = 3 , 2 , − 1 , the value 0 at x = 0 and the value 2 at x = 1 .
How did you derive f(x)= x[(x-1).h(x) + 2] ?
Log in to reply
If v = 1 then g ( 1 ) = 2 , and so g ( x ) = ( x − 1 ) h ( x ) + 2 .
Log in to reply
What is g(x) belongs to Z[x]?
Log in to reply
@Shaurya Gupta – Z [ x ] is the set of polynomials with integer coefficients.
Small typo. In numbered paragraph 1, below the table, the text should read h ( 2 ) = 1 , not 0 .
We can simplify this by applying the transformation
F ( x ) = 6 − f ( x )
Then we need to find the largest possible number of distinct roots of F , given that there must be values such that F ( n ) = 6 and F ( m ) = 4 .
Letting r i be the integer roots of F , we can write it as: F ( x ) = g ( x ) h ( x ) where g ( x ) = i ∏ ( x − r i ) s i and h ( x ) is the quotient F / g .
Now, h has no integer roots by definition; and since dividing a polynomial with integer coefficients by a factor x − r must leave a quotient with integer coefficients, h ( x ) must be a polynomial with integer coefficients; therefore h ( m ) and h ( n ) must be integers.
So, if F ( n ) = 6 then g ( n ) ∣ 6 , and similarly, g ( m ) ∣ 4 .
In order that g ( m ) ∣ 4 , the most number of possible distinct factors is four, and there's only one way of doing it: 2 × − 2 × 1 × − 1 , i.e.
g ( x ) = ( x − 2 ) ( x − 1 ) ( x + 1 ) ( x + 2 )
and g ( 0 ) = 4 . However, it can be readily checked that there is no solution to g ( n ) = 6 here, so we can rule out this case.
The next possible case is that g only has three factors. In this case,
g ( x ) = ( x + 2 ) ( x − 1 ) ( x − 2 )
gives g ( 0 ) = 4 and g ( − 1 ) = 6 .
Conveniently, h ( x ) = 1 will do here, and so we will get an f with 3 distinct integer roots.
It we are interested in finding an explicit form for f , then taking h = 1 : f ( x ) = 6 − ( x + 2 ) ( x − 1 ) ( x − 2 ) = − x 3 + x 2 + 4 x + 2
with f ( 0 ) = 2 , f ( − 1 ) = 0 , f ( − 2 ) = f ( 1 ) = f ( 2 ) = 6 .
The question could have been made a little trickier by requiring h to be nontrivial too, e.g. if we needed f ( m ) = − 2 2 instead , then we'd have to find f ( x ) = 6 − ( x + 2 ) ( x − 1 ) ( x − 2 ) ( 7 + 6 x ) or similar.
Log in to reply
True, but that would have been a bit too awkward to state, and it would not affect the main idea.
WLOG, after a change of variable, we can take f ( 0 ) = 0 and f ( a ) = 2 for an integer value of a > 0 . So we can write f ( x ) = x h ( x ) , so a ∣ 2 , so a = 1 or 2 .
If a = 1 , we can write f ( x ) = x ( ( x − 1 ) g 1 ( x ) + 2 ) . If a = 2 , we can write f ( x ) = x ( ( x − 2 ) g 2 ( x ) + 1 ) . Here the g i are polynomials with integer coefficients. In both cases f ( x ) = 6 can only be true for a ∣ 6 , so we only need to check x = ± 1 , ± 2 , ± 3 , ± 6 . There are seven values to check in each case (since 1 in case 1 and 2 in case 2 are already ruled out). Note that f ( x ) = 6 ⇔ g 1 ( x ) = x ( x − 1 ) 6 − 2 x in case 1, and g 2 ( x ) = x ( x − 2 ) 6 − x in case 2, so we first need to see for which values of x this is an integer.
So we get g 1 ( − 1 ) = 4 , g 1 ( 2 ) = 1 , g 1 ( − 2 ) = 5 / 3 , , g 1 ( 3 ) = 0 , g 1 ( − 3 ) = 1 , g 1 ( 6 ) = − 1 / 5 , g 1 ( − 6 ) = 3 / 7 . Throwing out the non-integer values, we see that there are at most four solutions, and the max is attained only if there is a polynomial g 1 with integer coefficients whose graph passes through ( − 1 , 4 ) , ( 2 , 1 ) , ( 3 , 0 ) , ( − 3 , 1 ) . But these last two are impossible since ( x − 3 ) would divide g 1 ( x ) and we'd get − 6 ∣ 1 . Hence the max in case 1 is at most 3 .
Similarly for g 2 , we get that there are at most four solutions, and the max is attained only if there is a polynomial g 2 with integer coefficients whose graph passes through ( − 1 , 5 ) , ( − 2 , 1 ) , ( − 3 , 1 ) , ( 6 , 0 ) . Again the last two points don't work because ( x − 6 ) would divide g 2 ( x ) and we'd get e.g. − 9 ∣ 1 . So the max in case 2 is at most 3 .
Finally, to show that the max is in fact 3 , we exhibit such a polynomial by setting g 1 ( x ) = 3 − x and solving to get f ( x ) = − x 3 + 4 x 2 − x . Here f ( 0 ) = 0 , f ( 1 ) = 2 , , and f ( x ) = 6 has the solutions x = − 1 , 2 , 3 .
Sorry, that ( − 1 , 5 ) in case 2 should actually be ( 1 , − 5 ) .
Knowing that: f ( n ) = 0 By the Polynomial remainder theorem we get that there is a polynomial Q ( x ) with integer coefficients such that f ( x ) can be written as
f ( x ) = Q ( x ) ( x − n )
Now we know that f ( m ) = 2 so
f ( m ) = Q ( m ) ( m − n ) = 2
And dividing by ( m − n ) we get
Q ( m ) = m − n 2
Q ( m ) has to be an integer, so m − n divides 2 , it is important that m − n is a constant, we will call it k , knowing that the only possible values for k are: 1 , 2 , − 1 and − 2 . Proceeding with calculations we obtain
Q ( m ) − k 2 = 0
Now we can consider the L H S as a polynomial in the variable m and we notice that the polynomial must have integer coefficients so (as before) there is a polyinomial R ( x ) with integer coefficients such that:
Q ( x ) − k 2 = R ( x ) ( x − m )
Q ( x ) = R ( x ) ( x − m ) + k 2
Substituting this in the alternative form of f ( x ) we found we obtain
f ( x ) = ( R ( x ) ( x − m ) + k 2 ) ( x − n )
Without loss of generality we can substitute x with X + n
f ( X + n ) = ( R ( X − n ) ( X − ( m − n ) ) + k 2 ) ( X )
we will call R ( X − n ) as R ′ ( X ) , and we notice that R ′ ( X ) is a polynomial with integer coefficients. We recall that k = m − n so
f ( X + n ) = ( R ′ ( X ) ( X − k ) + k 2 ) ( X )
The problem is equivalent to find for how many values of X the R H S is equal to 6 .
6 can be written as the product of two numbers (where the order matters) only in the following 8 ways:
We proceed with four cases, one for each possible value of k .
k = 2 we get: ( R ′ ( X ) ( X − 2 ) + 1 ) ( X ) = 6 we have that when X = 2 the other factor is 1 (so case 2 . has to be excluded), when X = − 1 the other factor must have reminder 1 when divided by 3 (case 5 . has to be excluded), in the same way when x = − 3 the other factor can not be − 2 (case 6 . has to be excluded) and when x = − 6 the other factor can not be − 1 (case 8 . has to be excluded). Therefore we are left with 4 cases. They can not be all true! In fact we have that for X = 6 , R ′ ( 6 ) must be 0 , so there is a polynomial with integer coefficients S ( X ) such that R ′ ( X ) = S ( X ) ( X − 6 ) Now substituting 3 in R ′ ( X ) we get that R ′ ( 3 ) must be a multiple of 3 . On the other hand ( R ′ ( X ) ( X − 2 ) + 1 ) can not be 2 in this case. We can say that at most there are 3 possible values for X .
k = 1 we get: ( R ′ ( X ) ( X − 1 ) + 2 ) ( X ) = 6 we have that when X = 1 the other factor is 2 (so case 1 . has to be excluded), when X = − 2 the other factor must have reminder 1 when divided by 3 (case 6 . has to be excluded), in the same way when x = 6 the other factor can not be 1 (case 4 . has to be excluded) and when x = − 6 the other factor can not be − 1 (case 8 . has to be excluded). Therefore we are left with 4 cases. It is easy to find a polynomial such that it works for X = − 1 , X = 2 and X = , in fact it is sufficient to show that R ′ ( − 1 ) has to be 4 , R ′ ( 2 ) has to be 1 and R ′ ( 3 ) has to be 0 . We can set R ′ ( X ) : = − X + 3 and we got a proper example. Using the Polynomial remainder theorem we have that all such polynomials must be of the form R ′ ( X ) = S ( X ) ( X + 1 ) ( X − 2 ) ( X − 3 ) + ( − X + 3 ) Where S ( X ) is a polynomial with integer coefficients. It is easy to see that there are no integer values of S ( 3 ) such that R ′ ( − 3 ) = 1 . So we can not have a solution for case 7 . .
k = − 2 we get the same as k = 2 changing the sign of X
k = − 1 we get the same as k = 1 changing the sign of X .
Therefore there are at most 3 solutions and they are given by the polynomial: f ( X − n ) = ( ( − X + 3 ) ( X − 1 ) + 2 ) ( X ) That is equivalent to: f ( x ) = ( ( − x + 3 + n ) ( x − n − 1 ) + 2 ) ( x − n )
How can I edit? That wasn't finished... f ( x ) = 6 iff x = n − 1 , x = n + 2 or x = n + 3 .
Log in to reply
You can't edit but you can continue with your solution.
Problem Loading...
Note Loading...
Set Loading...
I'm not entirely sure this is right, but here goes:
Consider g ( x ) = f ( x ) − 6 . It is a polynomial with integer coefficients which passes through − 6 and − 4 at integer values, and we want to maximize its number of roots.
We can shift our g ( x ) such that g ( 0 ) = − 4 . This means that the constant term of g ( x ) is − 4 , and that any rational roots must have their numerators be factors of 4 . This already caps the number of roots of g ( x ) at 6 (for ± 1 , ± 2 , and ± 4 ). More importantly, however, if we factor out all the integer roots r i of g ( x ) and write g ( x ) = ( x − r 1 ) ( x − r 2 ) … ( x − r n ) P ( x ) , then P ( x ) will have integer coefficients as well. This means that the constant c of P ( x ) is also an integer, so if − 4 = g ( 0 ) = ( − r 1 ) ( − r 2 ) … ( − r n ) c , then not only must each of the integer roots be a divisor of 4 but their product must be as well. This limits the number of integer roots to 4 (we could have ± 1 and ± 2 , for example, but we cannot take more than 4 different members of the group above without the product exceeding 4 ).
Finally, there is some integer x such that g ( x ) = − 6 . If we write: g ( x ) = a n x n + a n − 1 x n − 1 + … + a 1 x + a 0 , then the reasoning above told us that a 0 = − 4 . Thus, there is some x such that: a n x n + a n − 1 x n − 1 + … + a 1 x − 4 = − 6 , o r x ( a n x n − 1 + a n − 1 x n − 2 + … + a 1 ) = − 2
Hence, since the expression in parentheses is an integer polynomial and x is an integer as well, it must be the case that x is a divisor (positive or negative) of 2 . This means that x is an element of the set { ± 1 , ± 2 } , thus knocking out a prospective root from above.
This reduces the maximum number of roots to 3 . A quick check shows that, quite conveniently, g ( x ) = ( x + 2 ) ( x − 2 ) ( x + 1 ) has the value − 4 at x = 0 and − 6 at x = 1 . Thus, the maximum of 3 is indeed attainable.
Since the maximum number of roots of g ( x ) corresponds to the maximum number of times f ( x ) can be 6 , we know that the maximum is indeed 3 .
I'm not sure if there is a faster way.