Vertex-reinforced random walk on \(\mathbb Z\) eventually gets stuck on five points. (Q1889795): Difference between revisions

This paper contains the proof of the following conjecture of \textit{R. Pemantle} and \textit{S. Volkov} [Ann. Probab. 27, 1368--1388 (1999; Zbl 0960.60041)]. Consider the vertex-reinforced random walk on a locally finite graph (in this paper \(\mathbb Z\)), that is a random process which is more likely to visit vertices it has visited before. It was proved by Pemantle and Volkov that the range of this random walk is almost surely finite and that with positive probability the range contains exactly five points. Their conjecture is: With probability 1, the range contains exactly five points. The results and the heuristics (comparison with the Polya and Friedman urn model) developed by Pemantle and Volkov are used, but also some ideas from \textit{M. Benaïm} [in: Séminaire de probabilités XXXIII. Lect. Notes Math. 1709, 1--68 (1999; Zbl 0955.62085)].