Minimize the number of roots

Algebra Level 5

Suppose p ( x ) p(x) and q ( x ) q(x) are polynomials of degree 100 100 with complex coefficients, having no common zeroes. Find the smallest possible total number of complex zeroes of the polynomials p , p, q , q, and p q , p-q, counted without multiplicity.

Details and assumptions

When the zeroes are counted without multiplicity, the polynomial x 2 ( x 1 ) 3 x^2(x-1)^3 has two zeroes: x = 0 x=0 and x = 1. x=1.


The answer is 101.

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

Michael Tong
Aug 1, 2013

Note: This solution is somewhat incomplete or not as thorough as would be completely ideal. Unfortunately I must type this from a phone so I will give a general overview of the problem.

Clearly, the roots of p(x) and q(x) can be minimized when they have the least amount of roots themselves as possible. WLOG, say p ( x ) = ( x a ) 100 p(x) = (x - a)^{100} and q ( x ) = ( x b ) 100 q(x) = (x - b)^{100} , giving us roots a a and b b .

Then, we must find the minimum number of roots in p ( x ) q ( x ) p(x) - q(x) . This can be done by either getting rid of leading terms to decrease the degree or by having roots a or b.

Note that p ( x ) q ( x ) = 100 ( b a ) x 99 + ( 100 C 2 ) ( a 2 b 2 ) ) x 98 + . . . p(x) - q(x) = 100(b-a)x^{99} + (100 C 2)(a^2-b^2))x^{98} + ... (the degree was decreased by one because the first term of both polynomials is x 100 x^{100} ). By observing Vieta's formulas, it becomes clear that all of the roots of this polynomial must be distinct from each other and from a and b. Thus, there are a minimum of 1 + 1 + 99 = 101 1 + 1 + 99 = 101 roots.

Again, sorry for not explaining much, it's very hard to type a comprehensive solution on a phone (no internet here!). As soon as I can use my computer I will elaborate the solution in a reply.

Moderator note:

Minimizing the roots of p ( x ) p(x) and q ( x ) q(x) do not necessarily mean that we would minimize the total roots of p , q , p q p, q, p-q . There is no evidence that the total number cannot be less than 100.

The ABC conjecture has a proof that is currently being checked, and hasn't been verified as yet. This problem is related to Mason-Stothers theorem , which is the "polynomial analogue".

Now that I read the comments, I realize that some of the assumptions I make in the problem involve knowledge of the abc conjecture. http://en.wikipedia.org/wiki/Abc_conjecture

Michael Tong - 7 years, 10 months ago
Cody Johnson
Jul 28, 2013

Note that this solution is incomplete.

Let p ( x ) = ( x r 1 ) 100 p(x)=(x-r_1)^{100} and q ( x ) = ( x r 2 ) 100 q(x)=(x-r_2)^{100} . Since the x 100 x^{100} terms cancel, p ( x ) q ( x ) p(x)-q(x) must be of degree 99 (begin flaw) and therefore have 99 distinct roots (end flaw). This gives 1 + 1 + 99 = 101 1+1+99=\boxed{101} roots.

Moderator note:

This only shows that there are 2 polynomials which have a total of 101 complex zeros (given your flaw).

It needs to be shown that p , q , p 1 p, q, p-1 must have at least 101 complex zeros.

What is the proof that 101 101 indeed is the smallest possible number of zeroes? What if p ( x ) q ( x ) p(x) - q(x) has some repeated roots?

Sreejato Bhattacharya - 7 years, 10 months ago

Seems like the proof that the answer is at least 101 requires some form of the abc conjecture for polynomials, which is a theorem of Mason and Stothers. Here is an elementary proof: http://topologicalmusings.wordpress.com/2008/03/03/mason-stothers-theorem-and-the-abc-conjecture/

Patrick Corn - 7 years, 10 months ago

Log in to reply

Thanks for the link, Patrick.

There is one slight flaw in that proof -- namely, the way it invokes symmetry after having previously said "So, without any loss of generality, assume f and g are non-constant polynomials." But that's easily fixed: just note that if, say, g'=0 then f'g-fg'=f'g is nonzero as desired.

Peter Byers - 7 years, 10 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...