Markovian Navigability on Self-Similar Graph Approximations: Sub-Markov Semigroups, p-Adic Dynamics, and Pisot-Type Numeration
Year: 2026
Type: preprint
Abstract: We study a stochastic navigation model over path spaces associated with finite automata of Pisot-type substitutions. As a central example, we consider the Tribonacci substitution, whose prefix-suffix structure determines a directed graph with five transitions. We construct an injective topological encoding of admissible paths into the ring of $5$-adic integers and, independently, a contractive projection onto the complex geometric space associated with the substitution. On the graph, we introduce a discrete-time Markov chain with killing (absorption), governed by a substochastic transition matrix. For open symbolic targets, we define a hitting criterion based on the detection of cylinder sets contained within the target set. We prove that when the target is recognized by a deterministic finite automaton, the hitting probability prior to killing satisfies a finite linear system of harmonic equations. Furthermore, we study the approximation of open targets via finite unions of cylinders and establish conditions under which the corresponding hitting probabilities converge monotonically. Finally, we identify the analytical hypotheses required to extend the model to continuous-time processes and sub-Markovian semigroups on ultrametric spaces, providing an explicit structural example of fiber compatibility.
Author: Perdomo Leiva, Sebastian
Language: English
Subfield: Mathematical Physics
Field: Mathematics
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7236022891
