Skip to main content Accessibility help

We use cookies to distinguish you from other users and to provide you with a better experience on our websites. Close this message to accept cookies or find out how to manage your cookie settings.

Close cookie message
×
  1. Home
  2. Search

Refine search

Refine search

Access:
Content type:
Publication date:
Apply   From and To filters
Subject: Show more
Journals Show more
Publishers: Show more
Societies Show more
Series: Show more
Collections: Show more

Actions for selected content:

Select all | Deselect all
  • View selected items
  • Export citations
  • Download PDF (zip)
  • Save to Kindle
  • Save to Dropbox
  • Save to Google Drive

Save Search

You can save your searches here and later view and run them again in "My saved searches".

Please provide a title, maximum of 40 characters.
Cancel
×

90 results

Page 1 of 5

  • Surprise Probabilities in Markov Chains

  • Part of
  • JAMES NORRIS, YUVAL PERES, ALEX ZHAI
  • Journal:
    Combinatorics, Probability and Computing / Volume 26 / Issue 4 / July 2017
    Published online by Cambridge University Press:
    16 March 2017, pp. 603-627

  • In a Markov chain started at a state x, the hitting time τ(y) is the first time that the chain reaches another state y. We study the probability $\mathbb{P}_x(\tau(y) = t)$ that the first visit to y occurs precisely at a given time t. Informally speaking, the event that a new state is visited at a large time t may be considered a ‘surprise’. We prove the following three bounds.

    • • In any Markov chain with n states, $\mathbb{P}_x(\tau(y) = t) \le {n}/{t}$.

    • • In a reversible chain with n states, $\mathbb{P}_x(\tau(y) = t) \le {\sqrt{2n}}/{t}$ for $t \ge 4n + 4$.

    • • For random walk on a simple graph with n ≥ 2 vertices, $\mathbb{P}_x(\tau(y) = t) \le 4e \log(n)/t$.

    We construct examples showing that these bounds are close to optimal. The main feature of our bounds is that they require very little knowledge of the structure of the Markov chain.

    To prove the bound for random walk on graphs, we establish the following estimate conjectured by Aldous, Ding and Oveis-Gharan (private communication): for random walk on an n-vertex graph, for every initial vertex x,

    $$ \sum_y \biggl( \sup_{t \ge 0} p^t(x, y) \biggr) = O(\log n). $$

  • Chapter 6 - Isoperimetric Inequalities

  • Russell Lyons, Indiana University, Bloomington, Yuval Peres
  • Book:
    Probability on Trees and Networks
    Published online:
    19 January 2017
    Print publication:
    20 January 2017, pp 174-232

  • Summary

    Just as the branching number of a tree is for most purposes more important than the growth rate, there is a number for a general graph that is more important for many purposes than its growth rate. In the present chapter, we consider this number, or, rather, several variants of it, called isoperimetric or expansion constants. This is not an extension of the branching number, however; for that, the reader can see Section 13.7.

    Our main interest in expansion constants is to apply them to random walks and percolation on infinite graphs. In particular, whether the expansion is zero or positive plays a crucial role in determining qualitative behavior of these probabilistic processes. This will be seen here as well as in the later Chapters 7, 8, 10, and 11. A similar role is played on finite graphs, but we touch on finite graphs only briefly in Section 6.4.

    The second half of this chapter concerns isoperimetric and expansion profiles, functions that measure the size of the boundary of a set compared to the size of the set itself, when such a ratio is not necessarily bounded away from 0. Again, we present applications to random walks and to percolation.

    Flows and Submodularity

    A common illegal scheme for making money, known as a pyramid scheme or Ponzi scheme, goes essentially as follows. You convince 10 people to send you $100 each and to ask 10 others in turn to send them $100. Everyone who manages to do this will profit $900 (and you will profit $1000). Of course, some people will lose $100 in the end. But suppose that we had an infinite number of people. Then no one need lose money and indeed everyone can profit $900. But what if people can ask only people they know? Suppose that people are at the vertices of the square lattice and know only their four nearest neighbors. Is it now possible for everyone to profit $900? If the amount of money allowed to change hands (that is, the amount crossing any edge) is unbounded, then certainly this is possible. But what if the amount crossing any edge is bounded by, say, $1,000,000? The answer now is no.

Page 1 of 5