Skip to main content
12 events
when toggle format what by license comment
Sep 7, 2015 at 10:07 vote accept Flo140
Sep 7, 2015 at 9:35 vote accept Flo140
Sep 7, 2015 at 9:35
Sep 4, 2015 at 15:32 answer added Peter Mueller timeline score: 6
Sep 4, 2015 at 14:44 comment added Flo140 At first, i was studying some orthogonal function (in control theory) and more specifically its sign. The orthogonal function depends of four parameters and its linked with a polynomial (such as the one of the question). After a big number of numerical tests, i observe that when i take the parameters in the form $(abc,abd,acd,bcd)$ the polynomial always has a negative coefficient (that is important to understand the orthogonal function). I don't know if it's clear, it is always complicated to summarize a problem.
Sep 4, 2015 at 13:50 comment added Louis Deaett Yes, I certainly agree with you. My observation above is well short of answering the question, sadly. I'd also like to see how this could be resolved. I'd also be interested in knowing how this question came up to begin with, if it's at all easy to explain.
Sep 4, 2015 at 13:14 comment added Flo140 When we write $P(X)=\prod \phi_{m}(X)$ it's not obvious for me that we can deduce the sign of the coefficient of $P$. For exemple if we consider $$Q(X)=\frac{(1-X^{ab})(1-X^{ac})(1-X^{bc})}{(1-X^{a})(1-X^{b})(1-X^{c})}$$ we have also $Q(X)=\prod \phi_{m}(X)$ but in this case it's clear that $Q$ only has non negative coefficient.
Sep 4, 2015 at 12:42 comment added Flo140 @Louis the references you give seem really interesting...
Sep 3, 2015 at 22:46 comment added Louis Deaett Note that $P(x)=\prod \Phi_m(x)$, where the product is over all $1<m<abcd$ dividing $abcd$.
Sep 3, 2015 at 22:17 comment added Louis Deaett Related question I asked at mathoverflow.net/questions/214784/…
Sep 3, 2015 at 12:12 history edited Richard Stanley CC BY-SA 3.0
fixed typo
Sep 3, 2015 at 10:04 review First posts
Sep 3, 2015 at 10:19
Sep 3, 2015 at 9:56 history asked Flo140 CC BY-SA 3.0