20
$\begingroup$

Question P. Can a polynomial $P(x)=\sum_{n=0}^ma_nx^n$ with coefficients $a_n\in\{-1,1\}$ (and $P(1)=0$) have a multiple root in the interval $(\tfrac12,1)$?

Also I am interested in a similar question for analytic functions.

Question A. Let $f(x)=\sum_{n=0}^\infty a_nx^n$ a series with coefficients $a_n\in\{-1,1\}$ such that $\sup_{m\in\mathbb N}|\sum_{n=0}^ma_i|<\infty$. Can the analytic function $f$ have a multiple root in the interval $(0,1)$?

$\endgroup$
12
  • 5
    $\begingroup$ If we had an example in question P without extra condition $P(1)=0$, we can reach it by multiplying $P$ by $1-x^{m+1}$. Also this yields a positive answer to question $A$ by a similar reasoning $\endgroup$ Jun 10, 2022 at 12:18
  • 2
    $\begingroup$ Regarding P: even without the $P(1)=0$ restriction, up to $m=20$, there are no multiple roots at all for $m$ even, while for $m$ odd all multiple roots that may occur are $m+1$st roots of unity. $\endgroup$ Jun 10, 2022 at 16:02
  • 4
    $\begingroup$ Another thing about P: for each odd $m=2n-1$ there are two $P$'s with multiple roots, namely $\frac{(1-x^n)^2}{1-x}=1+x+...+x^{n-1}-x^n-x^{n+1}-...-x^{2n-1}$ and $\frac{(1-(-x)^n)^2}{1+x}=1-x+...+(-x)^{n-1}-(-x)^n-(-x)^{n+1}-...-x^{2n-1}$, and for $n$ prime these seem to be the only ones (again, checked experimentally up to $m=21$); but for $n$ not prime there are very many. Numbers of such polynomials go like $0,0,0,2,0,2,0,6,0,2,0,62,0,2,0,518,0,134,0,5452,0,2,...$ $\endgroup$ Jun 10, 2022 at 17:48
  • 4
    $\begingroup$ Such polynomials (series) are called Littlewood polynomials (series). Vader in Theorem 7.1 of “Real roots of Littlewood polynomials” showed that a Littlewood polynomial of even degree has no real roots. $\endgroup$ Jun 10, 2022 at 18:35
  • 3
    $\begingroup$ @TimothyChow, $(z^{18} + 2z^{15} + 2z^{13} + z^{12} + 2z^{11} + 3z^{10} + 3z^8 + 2z^7 + z^6 + 2z^5 + 2z^3 + z^2 + 1)(z^2 + 1)(z - 1)(z^3 - z - 1)^2$ and $(z^{18} + 2z^{15} + 2z^{13} + z^{12} + 3z^{10} + z^8 + 2z^7 + z^6 + 2z^5 + 2z^3 + z^2 + 1)(z^2 + 1)(z - 1)(z^3 - z - 1)^2$ are Littlewood polynomials with a non-cyclotomic repeated cubic factor. $\endgroup$ Jun 13, 2022 at 15:56

2 Answers 2

14
$\begingroup$

Question P. Can a polynomial $P(x)=\sum_{n=0}^ma_nx^n$ with coefficients $a_n\in\{-1,1\}$ (and $P(1)=0$) have a multiple root in the interval $(\tfrac12,1)$?

Yes. The following four Littlewood polynomials:

  • $z^{27} + z^{26} + z^{25} + z^{24} + z^{23} - z^{22} - z^{21} + z^{20} + z^{19} + z^{18} - z^{17} - z^{16} - z^{15} - z^{14} - z^{13} - z^{12} - z^{11} - z^{10} + z^9 + z^8 - z^7 + z^6 - z^5 + z^4 - z^3 + z^2 + z - 1 = (z^{18} + z^{16} + 2z^{15} + 2z^{13} + z^{12} + 2z^{11} + 3z^{10} + 3z^8 + 2z^7 + z^6 + 2z^5 + 2z^3 + 1)(z^2 + 1)(z - 1)(z^3 + z^2 - 1)^2$
  • $z^{27} + z^{26} + z^{25} - z^{24} - z^{23} - z^{22} + z^{21} - z^{20} - z^{19} + z^{18} + z^{17} - z^{16} - z^{15} + z^{14} + z^{13} - z^{12} - z^{11} - z^{10} - z^9 - z^8 + z^7 + z^6 - z^5 + z^4 - z^3 + z^2 + z - 1 = (z^{21} - z^{20} + 2z^{19} - 2z^{18} + z^{17} + z^{16} - 3z^{15} + 3z^{14} - 2z^{13} + 2z^{11} - 4z^{10} + 4z^9 - 2z^8 - z^7 + 3z^6 - 4z^5 + 2z^4 - z^3 - z^2 + z - 1)(z^3 + z^2 - 1)^2$
  • $z^{27} + z^{26} + z^{25} + z^{24} + z^{23} - z^{22} - z^{21} + z^{20} - z^{19} - z^{18} - z^{17} + z^{16} - z^{15} + z^{14} + z^{13} - z^{12} - z^{11} - z^{10} - z^9 - z^8 + z^7 + z^6 - z^5 + z^4 - z^3 + z^2 + z - 1 = (z^{18} + z^{16} + 2z^{15} + 2z^{13} + z^{12} + 2z^{11} + z^{10} + 3z^8 + z^6 + 2z^5 + 2z^3 + 1)(z^2 + 1)(z - 1)(z^3 + z^2 - 1)^2$
  • $z^{27} + z^{26} + z^{25} - z^{24} - z^{23} - z^{22} + z^{21} + z^{20} + z^{19} + z^{18} - z^{17} + z^{16} - z^{15} - z^{14} - z^{13} + z^{12} - z^{11} - z^{10} - z^9 - z^8 + z^7 + z^6 - z^5 + z^4 - z^3 + z^2 + z - 1 = (z^{18} + z^{16} + 2z^{13} - z^{12} + 2z^{11} + z^{10} + 3z^8 + z^6 + 2z^5 + 2z^3 + 1)(z^2 + 1)(z - 1)(z^3 + z^2 - 1)^2$

all have repeated factor $z^3 + z^2 - 1$ with root $z \approx 0.75488$.

$\endgroup$
4
  • 4
    $\begingroup$ Did you found them by exhaustive search or some other methods? Are these of smallest degree? $\endgroup$
    – Somnium
    Jun 13, 2022 at 20:43
  • 1
    $\begingroup$ @Somnium, exhaustive search, and yes. $\endgroup$ Jun 13, 2022 at 21:45
  • $\begingroup$ Wow. So mysterious! $\endgroup$ Jun 14, 2022 at 6:11
  • $\begingroup$ I've searched as far as I'm going to go. There are 16 Littlewood polynomials of degree 27 which have repeated non-cyclotomic roots: 4 each for 4 cubic repeated factors. There are none of degrees 29, 31, or 33, which I find mildly surprising. $\endgroup$ Jun 28, 2022 at 7:15
6
$\begingroup$

Not an answer, only some sort of evidence for A, hence community wiki. There seems to be an algorithm that produces a sequence of polynomials $(P_n)$ like in P, with $P_0=1$, $P_{n+1}=P_n\pm x^{n+1}$, $n\geqslant0$, and with $P_n$ having a local minimum at $m_n$ such that the sequence of $m_n$ converges to $\approx0.7257794$ while $P_n(m_n)$ tends to 0.

You do this: let $P_6=1-x-x^2-x^3+x^4+x^5+x^6$; it has a nice local minimum at $m_6\approx0.719842$. Now iterate the following procedure. Find minima of $P_n+x^{n+1}$, $P_n-x^{n+1}$ in the vicinity of $m_n$. Compare absolute values of these two polynomials at these minima. Choose for $P_{n+1}$ the one with smaller absolute value (and for $m_{n+1}$ where that value is attained).

Experimentally, these values become quite small quickly. For example, $P_{500}(m_{500})$ is about $1.644734\times 10^{-70}$.

Of course there is no control of the sequence $P_n(1)$ which is required to remain bounded. In fact $P_3(1)=-2$, $P_7(1)=2$, while $P_n(1)$ is either $-1$, $0$ or $1$ for all other $n$ up to $500$.

$\endgroup$
9
  • 2
    $\begingroup$ Is any periodicity in the sequence of pluses and minuses, found by this algorithm? $\endgroup$ Jun 11, 2022 at 4:28
  • 1
    $\begingroup$ @TarasBanakh None that I could detect. It goes like 1 plus, 3 minuses, 4 pluses, 2 minuses, 2 pluses, 2 minuses, 1 plus, $1, 1, 2, 1, 1, 2, 2, 1, 1, 2, 1, 1, 2, 2, 2, 1, 1, 2, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 2, 1,...$ $\endgroup$ Jun 11, 2022 at 4:33
  • 1
    $\begingroup$ @AlexRavsky At least it seems that forcing periodicity cannot work. Eventually periodic Littlewood series are of the form $P_n+x^{n+1}\frac{P_m}{1-x^{m+1}}$ where $P_n$ and $P_m$ are Littlewood polynomials of degrees $n$ and $m$ respectively. Experimentally, it seems that none of these have any multiple roots. $\endgroup$ Jun 11, 2022 at 13:18
  • 1
    $\begingroup$ @AlexRavsky Oh no sorry, there are such, starting from $m=7$. For example,$$1+x\frac{1-x-x^2-x^3-x^4+x^5+x^6+x^7}{1-x^8}=\frac{(1-x)(1+x)^2}{1+x^4}$$ $\endgroup$ Jun 11, 2022 at 14:13
  • 1
    $\begingroup$ @AlexRavsky I now checked up to $\max(m,n)\leqslant10$, the only double zeros are at roots of unity. Of course $10$ is not much, but still... $\endgroup$ Jun 12, 2022 at 7:59

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.