3
$\begingroup$

(This is a natural continuation of a previous post.)

I. Quintic method

Given the Lehmer quintic,

$$x^5 + n^2x^4 - (2n^3 + 6n^2 + 10n + 10)x^3 + (n^4 + 5n^3 + 11n^2 + 15n + 5)x^2 + (n^3 + 4n^2 + 10n + 10)x +1 = 0$$

From this post, we saw its roots $x_i$ can be ordered such that,

$$(x_1^4\, x_2^3\, x_3^2\, x_4)^{1/5} + (x_2^4\, x_3^3\, x_4^2\, x_5)^{1/5} + \dots + (x_5^4\, x_1^3\, x_2^2\, x_3)^{1/5} = 0$$

Or equivalently,

$$\frac1{x_1}-\frac1{x_1 x_2}+\frac1{x_1 x_2 x_3}-\frac1{x_1 x_2 x_3 x_4}+\frac1{x_1x_2x_3x_4x_5} = 0$$

This ordering is useful, since using the same order of roots, it turns out that,

$$(x_1\,x_2^4\,x_3^2\,x_4^7)^{1/11} + (x_2\,x_3^4\,x_4^2\,x_5^7)^{1/11} + \dots + (x_5\,x_1^4\,x_2^2\,x_3^7)^{1/11} = z_1$$ $$(x_1^2\,x_2\,x_3^7\,x_4^4)^{1/11} + (x_2^2\,x_3\,x_4^7\,x_5^4)^{1/11} + \dots + (x_5^2\,x_1\,x_2^7\,x_3^4)^{1/11} = z_2$$ $$(x_1^4\,x_2^7\,x_3\,x_4^2)^{1/11} + (x_2^4\,x_3^7\,x_4\,x_5^2)^{1/11} + \dots +(x_5^4\,x_1^7\,x_2\,x_3^2)^{1/11} = z_3$$ $$(x_1^7\,x_2^2\,x_3^4\,x_4)^{1/11} + (x_2^7\,x_3^2\,x_4^4\,x_5)^{1/11} + \dots +(x_5^7\,x_1^2\,x_2^4\,x_3)^{1/11} = z_4$$

where the $z_i$ are now roots of four different $11$-deg equations.

Notice that $z_1$ and $z_4$ are "complementary", with the $x_1, x_4$ of their starting terms just swapping exponents, likewise with the $x_2, x_3$. (The same can be said for $z_2$ and $z_3$).


II. Example

Let $n=-1$ and we have

$$x^5 + x^4 - 4x^3 - 3x^2 + 3x + 1 = 0$$

with roots $x_i = 2\cos\frac{2\pi k}{11}$ in the same order $k= 1,4,5,2,3.$ Using the four expressions above, these yield the four quintics,

\begin{align} -1 + 529 y + 5361 y^2 + 756 y^3 - 377 y^4 &+ y^5\\ -1 - 439 y - 45701 y^2 - 5536 y^3 - 135 y^4 &+ y^5\\ -1 - 4806 y - 5771 y^2 - 1543 y^3 - 47 y^4 &+ y^5\\ -1 + 408 y + 2215 y^2 + 1724 y^3 - 740 y^4 &+ y^5 \end{align}

such that,

$$y_1^{1/11}+y_2^{1/11}+y_3^{1/11}+y_4^{1/11}+y_5^{1/11} = z_i$$

are roots of the four $11$-degree equations,

\begin{align} -16954 - 2387 z + 3762 z^2 + 2200 z^3 + 704 z^4 & + 187 z^5 - 110 z^6 - 33 z^7 + 22 z^8 + z^{11}\\ 9908 + 8987 z + 7029 z^2 + 4499 z^3 + 1188 z^4 & - 418 z^5 - 352 z^6 - 33 z^7 + 22 z^8 + z^{11}\\ 7213 - 572 z - 1804 z^2 + 2563 z^3 + 1430 z^4 & - 418 z^5 - 352 z^6 - 33 z^7 + 22 z^8 + z^{11}\\ -18043 + 7777 z + 253 z^2 + 3168 z^3 - 385 z^4 & - 418 z^5 - 110 z^6 - 33 z^7 + 22 z^8 + z^{11} \end{align}

Incidentally, the unexplained phenomenon of shared coefficients (also found in the cubic method) appears again.


**III. Questions

  1. For the quintic with one root "fixed", there are now $(n-1)! = 24$ permutations, creating 24 different quintics. Why is it there are only four useful ones (I checked) that can solve an $11$-deg equation?
  2. And what would be one ($a,b,c,d)$ such that $(x_1^a\,x_2^b\,x_3^c\,x_4^d)^{1/p}$ and its Galois conjugates can be used to solve $p=31$?
$\endgroup$

1 Answer 1

3
+50
$\begingroup$

Question 1: An irreducible Lehmer quintic $L(n)$ has Galois group ${\bf Z} / 5 {\bf Z}$. Let $\sigma$ be a generator, and order the roots $x_i$ ($i \bmod 5$) so that $x_{i+1} = \sigma(x_i)$ for each $i$.

Note that $\prod_i x_i = -1$ because $L(n)$ has constant coefficient $1$. Thus $\prod_i x_i^{a_i} = (-1)^c \prod_i x_i^{a_i+c}$ for any integer $c$. You chose $c$ that makes $a_0 = 0$, but other choices may be more useful. For sums of (1/11)th powers, taking $c=6$ converts the exponents $(4,2,7,0,1)$ to $(10,8,2,6,7)$, proportional to the 5th roots of unity $\bmod 11$ (namely $c_i \equiv -3^i \bmod 11$ for $i=0,1,2,3,4$). The four choices of exponents correspond to the four isomorphisms from the Galois group to the group of 5th roots of unity $\bmod 11$.

Question 2: For (1/31)st powers we likewise need 5th roots of unity $\bmod 31$, which are $(1,2,4,8,16)$ (or ($0,1,3,7,15$) if you insist on having a zero exponent). Again there are four choices, again with "shared coefficients" including some zero coefficients, such as

x^31 + 744*x^26 - 620*x^25 - 4960*x^24 - 16058*x^23 + 27528*x^22 + 186372*x^21 + 13392*x^20 + 57908*x^19 - 2151400*x^18 + 6700526*x^17 + 6967374*x^16 + 33644796*x^15 - 86990712*x^14 - 125974204*x^13 - 244592976*x^12 + 559706116*x^11 + 1329731980*x^10 + 1937968348*x^9 + 760060790*x^8 - 615794292*x^7 - 539362428*x^6 + 376540942*x^5 + 474549116*x^4 + 105617372*x^3 - 15563860*x^2 - 3550678*x + 645413

(found numerically by computing to high precision and applying GP's aldgep).

$\endgroup$
2
  • $\begingroup$ Thanks! I found the exponents (4,2,7,0,1) through sheer trial-and-error, but I knew there had to be rhyme and reason why they worked. For prime $p$, is it safe to assume the number of automorphisms (hence the choices of exponents) for $p=7$ is 6? Or, in general, $p-1$? $\endgroup$ Jan 26, 2023 at 3:22
  • $\begingroup$ This MO question about your nonic might be of interest. $\endgroup$ May 1, 2023 at 21:01

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.