Challenging exercise about polynomials

Click For Summary
The discussion centers on a challenging exercise involving polynomials P and Q in the ring of integers, specifically proving that if a prime number p divides all coefficients of their product PQ, then p must divide the coefficients of either P or Q. Participants explore the implications of polynomial degrees and coefficients, emphasizing the need for rigorous proof and clarity in reasoning. Various approaches to the proof are suggested, including induction and the use of Bézout's identity. The conversation highlights the complexities of the proof and the necessity for detailed explanations to enhance understanding. Overall, the thread encourages collaborative problem-solving and deeper exploration of polynomial properties.
geoffrey159
Messages
535
Reaction score
72

Homework Statement



If ##P,Q## are polynomials of ##\mathbb{Z}[X]##, and ##p## is a prime number that divides all the coefficients of ##PQ##, show that ##p## divides the coefficients of ##P## or the coefficients of ##Q##.

Homework Equations



##c_n = \sum_{k=0}^{n} a_k b_{n-k} ## is the n-th coefficient of ##PQ##, where the ##(a_k)_{k=0..N}## are the coefficients of ##P##, and the ##(b_k)_{k=0..M}## are the coefficients of ##Q##

The Attempt at a Solution



That exercise gave me a serious headache, but it was an interesting one. Some people may want to try it for themselves.

@ homework helplers: Do you agree with my solution ?

I'll do it by induction.

  • Since ## p ## divides ## c_0 = a_0b_0 ##, and since ## p ## is prime, then ##p## divides ##a_0## or ##b_0##.

  • Assume that ##p|a_0, ..., a_n ## or ##p|b_0,...,b_n##.
    If ## n \ge \min\{\text{deg}(P),\text{deg}(Q)\} ## then there is nothing to prove, so we will assume that ## n < \min\{\text{deg}(P),\text{deg}(Q)\} ##. That last inequality and the fact that ##\text{deg}(PQ) = \text{deg}(P)+\text{deg}(Q) ## tell us that ##2n+2 \le \text{deg}(PQ) ##

Case 1: ##p|a_0, ..., a_n ## and there is ## k \in \{0...n\} ## such that ##p|b_0,...,b_{k-1}## and ##p\not | \ b_k ##.
Since ##p## divides ##c_{n+k+1} = (a_0b_{n+k+1} + ... + a_n b_{k+1}) + a_{n+1} b_k + (a_{n+2} b_{k-1} +...+a_{n+k-1}b_0)## , then ## p | a_{n+1}b_k ##, and with Gauss lemma, ##p|a_{n+1}##

Case 2: ##p|a_0, ..., a_n ## and ##p|b_0,...,b_n##.
Since ##p## divides ##c_{2n+2} = (a_0b_{2n+2} + ... + a_n b_{n+2}) + a_{n+1} b_{n+1} + (a_{n+2} b_{n} +...+a_{2n+2}b_0)##, then ## p | a_{n+1}b_{n+1} ##. Since ##p## is prime, it divides ##a_{n+1} ## or ##b_{n+1}##

Case 3 and 4: They are the symetric cases but assuming this time that ##p|b_0,...,b_n##
  • So by induction, ##p## divides the coefficients of ##P## or the coefficients of ##Q##
 
Physics news on Phys.org
geoffrey159 said:

Homework Statement



If ##P,Q## are polynomials of ##\mathbb{Z}[X]##, and ##p## is a prime number that divides all the coefficients of ##PQ##, show that ##p## divides the coefficients of ##P## or the coefficients of ##Q##.

Homework Equations



##c_n = \sum_{k=0}^{n} a_k b_{n-k} ## is the n-th coefficient of ##PQ##, where the ##(a_k)_{k=0..N}## are the coefficients of ##P##, and the ##(b_k)_{k=0..M}## are the coefficients of ##Q##

The Attempt at a Solution



That exercise gave me a serious headache, but it was an interesting one. Some people may want to try it for themselves.

@ homework helplers: Do you agree with my solution ?

I'll do it by induction.

  • Since ## p ## divides ## c_0 = a_0b_0 ##, and since ## p ## is prime, then ##p## divides ##a_0## or ##b_0##.

  • Assume that ##p|a_0, ..., a_n ## or ##p|b_0,...,b_n##.
    If ## n \ge \min\{\text{deg}(P),\text{deg}(Q)\} ## then there is nothing to prove, so we will assume that ## n < \min\{\text{deg}(P),\text{deg}(Q)\} ##. That last inequality and the fact that ##\text{deg}(PQ) = \text{deg}(P)+\text{deg}(Q) ## tell us that ##2n+2 \le \text{deg}(PQ) ##

Case 1: ##p|a_0, ..., a_n ## and there is ## k \in \{0...n\} ## such that ##p|b_0,...,b_{k-1}## and ##p\not | \ b_k ##.
Since ##p## divides ##c_{n+k+1} = (a_0b_{n+k+1} + ... + a_n b_{k+1}) + a_{n+1} b_k + (a_{n+2} b_{k-1} +...+a_{n+k-1}b_0)## , then ## p | a_{n+1}b_k ##, and with Gauss lemma, ##p|a_{n+1}##

Case 2: ##p|a_0, ..., a_n ## and ##p|b_0,...,b_n##.
Since ##p## divides ##c_{2n+2} = (a_0b_{2n+2} + ... + a_n b_{n+2}) + a_{n+1} b_{n+1} + (a_{n+2} b_{n} +...+a_{2n+2}b_0)##, then ## p | a_{n+1}b_{n+1} ##. Since ##p## is prime, it divides ##a_{n+1} ## or ##b_{n+1}##

Case 3 and 4: They are the symetric cases but assuming this time that ##p|b_0,...,b_n##
  • So by induction, ##p## divides the coefficients of ##P## or the coefficients of ##Q##

I really think you could elaborate more here:

If ## n \ge \min\{\text{deg}(P),\text{deg}(Q)\} ## then there is nothing to prove, so we will assume that ## n < \min\{\text{deg}(P),\text{deg}(Q)\} ##. That last inequality and the fact that ##\text{deg}(PQ) = \text{deg}(P)+\text{deg}(Q) ## tell us that ##2n+2 \le \text{deg}(PQ) ##

You seem to jump without explaining much detail. Perhaps show there is nothing to prove where you claim, and show some additional inequalities.

If I may offer an alternative argument, there is an easier way to express the requirement. I will provide an intro to the proof if you are interested:

Supposing ##p## is a prime number that divides all the coefficients of ##PQ##, we wish to show ##p## divides the coefficients of ##P## or the coefficients of ##Q##.

Please note: When I write something like ##p|Q## below, I am strictly speaking about dividing the coefficients.

Suppose ##p|Q##, but ##p \nmid P##. We must show ##p|Q##.

Since ##p \nmid P##, ##\exists s, t \in \mathbb{Z} \space | \space 1 = \cdots##.

The proof is almost complete, but there would be no fun in just seeing the answer.
 
Zondrina said:
I really think you could elaborate more here ...
You seem to jump without explaining much detail. Perhaps show there is nothing to prove where you claim, and show some additional inequalities.
Hi, thanks for replying !

I meant that if ## n \ge \min\{\text{deg}(P),\text{deg}(Q)\}##, and with the induction hypothesis, ##p## divides by definition all the coefficients of ##P## or ##Q##. The interesting part is when ## n < \min\{\text{deg}(P),\text{deg}(Q)\} ##.

If ## n < \min\{\text{deg}(P),\text{deg}(Q)\} ## then ##\text{deg}(PQ) = \text{deg}(P)+\text{deg}(Q) \ge 2 \min\{\text{deg}(P),\text{deg}(Q)\} \ge 2 (n+1) = 2n +2 ##, and since ##p## divides all the coefficients of ##PQ##, ##p | c_{0},...,c_{2n+2}, ..., c_{\text{deg}(PQ)}##
Zondrina said:
If I may offer an alternative argument, there is an easier way to express the requirement. I will provide an intro to the proof if you are interested:
...
Suppose ##p|Q##, but ##p \nmid P##. We must show ##p|Q##.
Since ##p \nmid P##, ##\exists s, t \in \mathbb{Z} \space | \space 1 = \cdots##.
The proof is almost complete, but there would be no fun in just seeing the answer.

I was almost sure there was some easier proof, but did not manage to find it. Please explain.
 
geoffrey159 said:
Hi, thanks for replying !

I meant that if ## n \ge \min\{\text{deg}(P),\text{deg}(Q)\}##, and with the induction hypothesis, ##p## divides by definition all the coefficients of ##P## or ##Q##. The interesting part is when ## n < \min\{\text{deg}(P),\text{deg}(Q)\} ##.

If ## n < \min\{\text{deg}(P),\text{deg}(Q)\} ## then ##\text{deg}(PQ) = \text{deg}(P)+\text{deg}(Q) \ge 2 \min\{\text{deg}(P),\text{deg}(Q)\} \ge 2 (n+1) = 2n +2 ##, and since ##p## divides all the coefficients of ##PQ##, ##p | c_{0},...,c_{2n+2}, ..., c_{\text{deg}(PQ)}##

I was almost sure there was some easier proof, but did not manage to find it. Please explain.

Suppose ##p|Q##, but ##p \nmid P##. We must show ##p|Q##.

Since ##p \nmid P##, ##\exists s, t \in \mathbb{Z} \space | \space 1 = Ps + pt \Rightarrow Q = PQs + pQt##.

Now, ##p## divides the right hand side, so ##p## also divides ##Q##.

Q.E.D.
 
Zondrina said:
You seem to jump without explaining much detail ...

That works for you too !I don't understand the notations that are used. What I get is the following

You say ##p| c_0,...,c_{\text{deg}(P)+\text{deg}(Q)} ## and you assume ##p\not | a_0,...,a_{\text{deg}(P)} ##, and you want to show ##p|b_0,...,b_{\text{deg}(Q)} ##
So what I understand is :
There is some ##a_k## such that ##\text{gcd}(p,a_k) = 1##, so that the family ##(a_0,...,a_{\text{deg}(P)},p)## is relatively prime.
And now I understand you use Bezout theorem. You say there are ##s_0,...,s_{\text{deg}(P)},t## in ##\mathbb{Z}## such that
## 1 = a_0 s_0 + ... +a_{\text{deg}(P)}s_{\text{deg}(P)} + tp ##

What's the next step?
 
Question: A clock's minute hand has length 4 and its hour hand has length 3. What is the distance between the tips at the moment when it is increasing most rapidly?(Putnam Exam Question) Answer: Making assumption that both the hands moves at constant angular velocities, the answer is ## \sqrt{7} .## But don't you think this assumption is somewhat doubtful and wrong?

Similar threads

  • · Replies 3 ·
Replies
3
Views
1K
  • · Replies 13 ·
Replies
13
Views
2K
  • · Replies 13 ·
Replies
13
Views
2K
  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 13 ·
Replies
13
Views
2K
  • · Replies 1 ·
Replies
1
Views
1K
  • · Replies 2 ·
Replies
2
Views
2K
  • · Replies 3 ·
Replies
3
Views
2K
Replies
3
Views
1K
Replies
3
Views
3K