Random walk on a finite group

From MaRDI portal





The paper gives a radically new approach to an [apparently!] straight forward question: A random traveler visits the \(\sigma\) vertices of a regular polytope \({\mathcal P}\) in the usual way (the walk is Markovian from vertex to vertex and gives to every vertex the same chance of being visited from the preceding one). A usual problem is to determine p(n), the probability of coming back to the initial position at the end of the n-th step. The main feature here is that you have to handle with the n-th power of a \(\sigma\times \sigma\) matrix and, even in simple cases, the so called theoretical solution proves to be a poor one. (Example: 120-cell polytope; \(\sigma =600!.)\) The author gives a fairly general way to reduce drastically the effective calculations to get the exact value of p(n). He introduces a new Markov chain related to the given one but with a striking smaller number of states (example: 120-cell polytope; new state space of order \(m=44!)\). By the way, the given method, not only reduces considerably the effective calculations but offers a deep excursion into the theory of representations of finite groups; moreover, the author announces a forthcoming paper giving examples for application of the given method.











This page was built for publication: Random walk on a finite group

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3317845)