I'm a math student (last year undergraduate) and I really enjoy solving math problems but sometimes I face problems that I don't see how to solve even after trying every idea I had. When I look up at the solution I understand it perfectly but I can't figure out how the author actually found his solution. Their thinking process is never explained. Let me take an example:
Let $P \in \Bbb R[X]$. Show that if $\forall x \in [0,1], P(x) \ge 0$ then $\exists A,B,C,D \in \Bbb R[X], P=A^2+XB^2+(1-X)C^2+X(1-X)D^2$.
What I've done:
- I tried to look at some special cases ($\deg P = 0,1,2$).
- tried to see if there is a structure behind this (by introducing $E= \{A^2+XB^2+(1-X)C^2+X(1-X)D^2 \ | \ A,B,C,D \in \Bbb R[X] \}$ and looking if $E$ is stable by product) but this was hard.
- tried to reinforce the hypothesis by assuming $P \ge 0$, this worked well because I remembered another problem I solved ($P\ge 0 \implies \exists A,B \in \Bbb R[X], P=A^2 + XB^2$). I tried to adapt the proof (what I already tried previously) but it didn't help.
- tried to use coefficients (really bad idea here).
- tried Lagrange interpolation (not helpful either).
I posted this problem here and @orangeskid advised me to show that, in fact, $P(x)= A^2(x) + x(1-x)D^2(x)$ ie. we could let $B=C=0$, it was a very good hint. Even in a parallel world I'd never have thought of doing so. But why didn't I think of doing this? I don't know, it wasn't obvious for me or maybe I don't see far enough.
Another example:
Let $A, B \in \Bbb Z[X]$ such that $(\gcd(A(n), B(n)))_n$ is periodic. What can we say about $A$ and $B$?
What I tried:
- I supposed that $\deg A = \deg B = 1$ even with that $\gcd(A(n),B(n))$ is something that scares me because I don't have a lot of intuition on this sequence. So I applied, Bézout's theorem which says that we can find two sequences $a_n,b_n$ such that $(a_n A(n) + b_n B(n))_n$ is periodic. And I got stuck here.
I saw a solution here using the resultant of two polynomials, something I never heard of.
There are other examples here, here and here: they all need a little trick, something close to magic (is that cleverness?).
So my questions are how could I find these tricks myself and make them natural? Why the origin of these tricks isn't explained? Should I just learn them "on the job"? Should I be worried (as a student in mathematics) if I don't see them? Thanks for your time!