Heaps and two exponential structures

From MaRDI portal



Abstract: Take sfQ=(sfQ1,sfQ2,ldots) to be an exponential structure and M(n) to be the number of minimal elements of sfQn where M(0)=1. Then a sequence of numbers rn(sfQn)nge1 is defined by the equation �egin{eqnarray*} sum_{nge 1}r_n({sf Q}_n)frac{z^n}{n!,M(n)}=-log(sum_{nge 0}(-1)^nfrac{z^n}{n!,M(n)}). end{eqnarray*} Let denote the poset sfQn with a hat0 adjoined and let hat1 denote the unique maximal element in the poset sfQn. Furthermore, let musfQn be the M"{o}bius function on the poset . Stanley proved that rn(sfQn)=(−1)nmusfQn(hat0,hat1). This implies that the numbers rn(sfQn) are integers. In this paper, we study the cases sfQn=Pin(r) and sfQn=sfQn(r) where Pin(r) and sfQn(r) are posets, respectively, of set partitions of [rn] whose block sizes are divisible by r and of r-partitions of [n]. In both cases we prove that rn(Pin(r)) and rn(sfQn(r)) enumerate the pyramids by applying the Cartier-Foata monoid identity and further prove that rn(Pin(r)) is the generalized Euler number Ern−1 and that rn(sfQn(2)) is the number of complete non-ambiguous trees of size 2n−1 by bijections. This gives a new proof of Welker's theorem that rn(Pin(r))=Ern−1 and implies the construction of r-dimensional complete non-ambiguous trees. As a bonus of applying the theory of heaps, we establish a bijection between the set of complete non-ambiguous forests and the set of pairs of permutations with no common rise. This answers an open question raised by Aval {it et al.}.












This page was built for publication: Heaps and two exponential structures

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5964258)