Let $f∈Q[X]$ and not constant or of the form $(x−a)^n$. Suppose:
$f_1:=\frac{f}{gcd(f,D^2f)}$
and
$f_2:=\frac{f_1}{gcd(f_1,Df_1)}$
where $Df$ stands for the formal derivative.
Is it true that $gcd(f_2,Df_2)=gcd(f_2,D^2f_2)=1$ ?
|
0
1
|
Let $f∈Q[X]$ and not constant or of the form $(x−a)^n$. Suppose: $f_1:=\frac{f}{gcd(f,D^2f)}$ and $f_2:=\frac{f_1}{gcd(f_1,Df_1)}$ where $Df$ stands for the formal derivative. Is it true that $gcd(f_2,Df_2)=gcd(f_2,D^2f_2)=1$ ? |
||
|
|
|
4
|
No. Let $$f(x)=(x-3x^3)(x+1/3).$$ Then $f''(-1/3)=0$ and $$f_1(x)=x-3x^3.$$ Since $f_1$ and $f_1'$ have no common roots, $f_2=f_1$. But $gcd(f_2,D^2f_2)=x$. |
||
|
|
|
0
|
No. Consider polynomials which are also odd functions. Edit: The above may be useful, but is not correct. If $f_2$ were odd, the answer might be no, but it is unclear that the relations above can ever produce an odd function. End Edit. Edit2: Thanks to Michael Renardy for producing an odd $f_2$. End Edit2. Gerhard "Ask Me About System Design" Paseman, 2011.04.12 |
|||
|
|