COUNTING QUADRANT WALKS WITH LARGE STEPS
The enumeration of quadrant walks with small steps (that is, steps taken in some of the 8 directions N, S, E, W, NE, NW, SE, SW) is now well understood.
In particular, their generating function is D-finite (i.e., solution of a linear differential equation)
if and only if a certain group of rational transformation, associated with the set of allowed steps, is finite.
It is far from obvious to extend the methods that led to this classification to quadrant walks with arbitrary steps.
Fayolle and Raschel have already discussed the difficulties that one expects using the complex analysis approach that was very powerful in the small step case.
In this talk, I will describe some progresses in the formal series approach, whose main ambition is to yield results when the group is finite.
(joint work with Alin Bostan and Steve Melczer)
MULTIDIMENSIONAL REFLECTED RANDOM WALK--SOME RESULTS AND MANY QUESTIONS
Let $\mu$ be a probability measure on $\mathbb{R}^{r+s}$ and
let $(Y_n,V_n)$ be i.i.d. $\mu$-distributed with $Y_n \in \mathbb{R}^r$
and $V_n \in \mathbb{R}^s$. Reflected random walk starting at $x \in \mathbb{R_+}^r$
is defined recursively by
$
X_0^x=x\,, \quad X_n^x = |X_{n-1}^x - Y_n|\,, \quad \text{where}\quad
|(a_1, \dots, a_r)| = (|a_1|, \dots, |a_r|)\,.
$
In $\mathbb{R}^s$, consider the ordinary sum $S_n=V_1 + \dots + V_n\,$.
We are interested in (topological) recurrence of the process $(X_n^x,v+S_n)$
starting at $(x,v)$.
While this is quite well understood for refelcted random walk in dimension $1$,
in higher dimension ($r \ge 2$) or with some non-reflected coordinates
($s \in \{ 1,2\}$) we have a few basic results and various open questions
with some partial answers.
This is ongoing work with Judith Kloas, with input from Marc Peigné and
Wojciech Cygan.