arXiv++ Combinatorics

We study the trinomial equation $x^n +px +q =0$. Here $p$ and $q$ are both real and nonzero. For $n\ge3$, expressions for the roots have been published as hypergeometric series in powers of the parameter $q^{n-1}/p^n$. For the special case of the cubic ($n=3$), we employ Kummer's identities to derive alternative series solutions in powers of the discriminant $D$, and also series in powers of $1/D$. We next derive new series, in powers of $D$ and also in powers of $1/D$, for all $n\ge 3$. The resulting series suggest identities analogous to Kummer's identities, for higher order hypergeometric series.

We treat success runs of independent identically distributed Bernoulli trials (with success parameter $p$) distributed according to the Type II binomial distribution of order $k$. However, the success runs are separated by a gap $g\ge1$ (a failure followed by $g-1$ arbitrary outcomes). Most of the literature treats the case $g=1$ only. Our main results are expressions for the probability mass function (we present two derivations) and the distribution of the longest success run. We also present more concise expressions for previously published results for the factorial moments. We present results for the mean, variance, probability mass function and factorial moments for $\textrm{NB}_{\rm II}(k,g,r)$, the Type II negative binomial distribution of order $k$, where the number of success runs $r$ is fixed and the number of trials $n$ is variable. Let $L$ denote the length of the longest success run. We present a recurrence and generating function for the distribution of $L$ and derive expressions for the mean, variance and factorial moments of $L$.

The recurrence for the $k$-generalized Fibonacci polynomials is usually iterated upwards to positive values of $n$ only. When the recurrence is iterated downwards to $n<0$, there are indices where the polynomials vanish identically. This fact does not seem to have been noted in the literature. We derive the set of such indices. We present the left-justified generalized Pascal triangle for $n<0$. For $k\ge3$ and $n<0$, we show that the degree of the polynomial does not increase monotonically with $|n|$. We derive expressions for the individual polynomial coefficients (the elementary symmetric polynomials of the roots). We present results for the properties of the polynomials, for both $n>0$ and $n<0$, including factorization of the polynomials and properties of the roots (including bounds on the amplitudes of the nonzero roots). Results are also derived for real roots. (Separate treatments are required for $n>0$ and $n<0$.) We employ generating functions to derive new combinatorial sums for the polynomials. The sums are more concise and computationally more efficient than previously published expressions. We also exhibit the relation of the $k$-generalized Jacobsthal and Pell polynomials to the Fibonacci polynomials.

We present a general formula for the particular solution of an inhomogeneous linear difference equation with variable coefficients. The answer is expressed as a weighted sum of fundamental solutions of the associated linear difference equation. This corresponds to an initial value problem in the case of linear differential equations. We remark that Green's functions are naturally suited for solving such problems. This note presents a Green's function formalism to solve an inhomogeneous linear difference equation with variable coefficients. Both the retarded and advanced Green's functions are required, to obtain a complete solution. We independently confirm previous work for the case of linear difference equations with constant coefficients.

We present new expressions for the $k$-generalized Fibonacci numbers, say $F_k(n)$. They satisfy the recurrence $F_k(n) = F_k(n-1) +\dots+F_k(n-k)$. Explicit expressions for the roots of the auxiliary (or characteristic) polynomial are presented, using Fuss-Catalan numbers. Properties of the roots are enumerated. We quantify the accuracy of asymptotic approximations for $F_k(n)$ for $n\gg1$. Our results subsume and extend some results published by previous authors. We also present a basis (or `fundamental solutions') to solve the above recurrence for arbitrary initial conditions. We comment on the use of generating functions and multinomial sums for the $k$-generalized Fibonacci numbers and related sequences. We note that the resulting multinomial sums are Dickson polynomials of the second kind in several variables. We also present what may be a new identity for companion matrices.

The negative binomial distribution NB$(k,r)$ of Type I is the probability distribution for a sequence of independent Bernoulli trials (with success parameter $p\in(0,1)$) with $r$ nonoverlapping success runs of length $\ge k$. We present a new, more concise, expression for its probability mass function. We show it can also be succinctly written using hypergeometric functions. We also present new expressions (combinatorial sums) for its moments and factorial moments, as opposed to only the mean and variance (which are already known). Next, we present an alternative non-combinatorial viewpoint, which yields expressions for the factorial moments not only for nonoverlapping success runs, but also for runs with an overlap of $\ell$, where $\ell\in[0,k-1]$. The case $\ell=k-1$ is the negative binomial distribution NB$(k,r)$ of Type III. The results also yield the solution for the negative binomial distribution NB$(k,r)$ with a minimum gap between the success runs (explained in the text). Addendum 1/23/2024: The probability mass function and factorial moments are derived from the probability generating function. Addendum 1/26/2024: Alternative expressions are presented for the negative binomial distribution NB$(k,r)$ of Type II.

We derive a simple expression for the $r^{th}$ factorial moment $μ_{(r)}$ of the geometric distribution of order $k$ with success parameter $p\in(0,1)$ (and $q=1-p$) in terms of its probability mass function $f_k(n)$. Specifically, $μ_{(r)} = r!f_k((r+1)k+r)/((qp^k)^{r+1})$.

The factorial moments of the standard Poisson distribution are well known and are simple, but the raw moments are considered to be more complicated (Touchard polynomials). The present note presents a recurrence relation and an explicit combinatorial sum for the raw moments of the Poisson distribution of order $k$. Unlike the standard Poisson distribution (the case $k=1$), for $k>1$ the structure of the raw and factorial moments have many similarities and the raw moments are not more complicated (formally, at least) than the factorial moments. We remark briefly on the central moments (i.e.~moments centered on the mean) of the Poisson distribution of order $k$.

The factorial moments of the standard Poisson distribution are well known. The present note presents an explicit combinatorial sum for the factorial moments of the Poisson distribution of order $k$. Unlike the standard Poisson distribution (the case $k=1$), for $k>1$ the structure of the factorial moments is much more complicated. Some properties of the factorial moments of the Poisson distribution of order $k$ are elucidated in this note.

Kostadinova and Minkova published an expression for the probability mass function (pmf) of the Poisson distribution of order $k$, as a combinatorial sum ($\mathit{Pliska~Stud.~Math.~Bulgar.}\ {\bf 22},\ 117-128\ (2013)$). Inspired by their elegant solution, this note presents an alternative combinatorial sum for the pmf of the Poisson distribution of order $k$. The terms are partitioned into blocks of length $k$ (as opposed to $k+1$ by Kostadinova and Minkova). The new sum offers an advantage in the following sense. For $n\in[rk+1,(r+1)k]$, the lowest power of $λ$ in the pmf is $λ^{r+1}$. Hence the lower limit of summation can be increased, to avoid needlessly calculating terms which cancel to identically zero.

We present an exposition on the Fuss--Catalan numbers, which are a generalization of the well known Catalan numbers. The literature on the subject is scattered (especially for the case of multiple independent parameters, as will be explained in the text), with overlapping definitions by different authors and duplication of proofs. This paper collects the main theorems and identities, with a consistent notation. Contact is made with the works of numerous authors, including the early works of Lambert and Euler. We demonstrate the application of the formalism to solve algebraic equations by infinite series. Our main result in this context is a new necessary and sufficient formula for the domain of absolute convergence of the series solutions of algebraic equations, which corrects and extends previous work in the field. Some historical material is placed in an Appendix.