The \(k\)-extension of a Mahonian statistic (Q1902815)

From MaRDI portal
scientific article
Language Label Description Also known as
English
The \(k\)-extension of a Mahonian statistic
scientific article

    Statements

    The \(k\)-extension of a Mahonian statistic (English)
    0 references
    0 references
    30 May 1996
    0 references
    From the author's paper: Let \(r\), \(l\), \(k\) be three non-negative integers such that \(1\leq r\) and \(l+ k= r\). Let \(X\) denote the alphabet \([r]= \{1, 2,\dots, r\}\) equipped with the natural order of the integers. Call the letters \(1, 2,\dots, l\) small and the letters \(l+ 1,l+ 2,\dots, l+k= r\) large. Also let \(S= S_k= \{1, 2,\dots, l\}\) and \(L= L_k= \{l+ 1, l+2,\dots,r\}\). Let \(w= x_1 x_2\cdots x_m\) be a word in the alphabet \(X\) and denoted by \(\overline w= y_1 y_2\cdots y_m\) its non-decreasing rearrangement. Let \(1\leq i\leq m\); say that \(i\) is a place of \(k\)-descent (resp. place of \(k\)-exceedance) for \(w\), if \(i\neq m\) and \(x_i> x_{i+ 1}\) or \(x_i= x_{i+ 1}\geq l+ 1\), or if \(i= m\) and \(x_i\geq l+ 1\) (resp. if \(x_i> y_i\) or \(x_i= y_i\geq l+ 1\)). The number of places of \(k\)-descent (resp. of places of \(k\)-exceedance) in \(w\) is denoted by ``\(\text{des}_k w\)'' (resp. ``\(\text{exc}_k w\)''). \textit{R. J. Clarke} and \textit{D. Foata} [Eur. J. Comb. 15, No. 4, 345-362 (1994; Zbl 0811.05069)] proved that ``\(\text{des}_k\)''and ``\(\text{exc}_k\)'' were equidistributed on each rearrangement class \(R({\mathbf c})\). In their second paper [\((*)\) Eur. J. Comb. 16, No. 3, 221-252 (1995; Zbl 0822.05066)] they constructed a bijection \(\rho\) of \(R({\mathbf c})\) onto itself satisfying \((\text{des}_k, \text{maj}_k)(w)= (\text{exc}_k, \text{den}_k)(\rho(w))\), where ``\(\text{maj}_k\)'' and ``\(\text{den}_k\)'' are two \(k\)-extensions of Mahonian statistics in the following sense. Let \(\text{SPD}_k w\) (resp. \(\text{SPE}_kw\)) be the sum of the places of \(k\)-descents (resp. of \(k\)-exceedances) in \(w\). \textit{R. J. Clarke} and \textit{D. Foata} [see \((*)\)] defined the \(k\)-major index, ``\(\text{maj}_k w\)'', of \(w\) as \(\text{maj}_k w= \text{SPD}_k w\). It would have seemed natural to define ``\(\text{den}_k w\)'' as \(\text{den}_k w= \text{SPE}_k w\), but \(\text{SPD}_k\) is not equidistributed with \(\text{maj}_k\) on each class \(R({\mathbf c})\). However, the inequality \(\text{den}_k w\geq \text{SPE}_k w\) holds. Therefore, the natural question arises: can well-defined predicates be found to characterize the (non-negative) difference ``\(\text{den}_k- \text{SPE}_k\)'' and then provide the ``natural'' definition of ``\(\text{den}_k\)''? For \(k= 0\) the question had been answered successfully by the author. It is the purpose of this paper to provide the predicates for an arbitrary \(k\).
    0 references
    0 references
    Denert statistics
    0 references
    Eulerian calculus
    0 references
    descent
    0 references
    exceedance
    0 references
    alphabet
    0 references
    word
    0 references
    rearrangement
    0 references
    Mahonian statistics
    0 references
    0 references
    0 references