Title Information
Title
Applications of randomized algorithms to counting problems
Name: Personal
Name Part
Sukurdeep, Yashil
Role
Role Term: Text
creator
Name: Personal
Name Part
Dupuis, Paul
Role
Role Term: Text
advisor
affiliation
Brown University. Department of Applied Mathematics
Name: Corporate
Name Part
Brown University. Karen T. Romer Undergraduate Teaching and Research Awards
Role
Role Term: Text
research program
Type of Resource
text
Genre (aat)
posters
Origin Information
Place
Place Term: Code (MARC Country Code)
riu
Place
Place Term: Text
Providence, RI
Publisher
Brown University
Date Created (keyDate="yes", encoding="w3cdtf")
2017
Physical Description
Extent
1 poster
digitalOrigin
reformatted digital
Abstract
We study the binary contingency table problem, where our goal is to count the number of n x m binary tables ({0,1}-valued matrices) that satisfy certain given row and column sums. We present a straightforward Markov Chain Monte Carlo (MCMC) algorithm that gives robust estimates for the number of binary contingency tables when the dimension of the matrices is relatively low. We then present the parallel tempering method, which makes use of coupled Markov Chains running at different "temperatures", for approximately counting the number of binary contingency tables. We then discuss the qualitative properties of the parallel tempering method and its advantages with regards to other randomized algorithms such as the splitting algorithm.
Subject (LCSH)
Topic
Markov processes
Subject (LCSH)
Topic
Monte Carlo method
Subject (LCSH)
Topic
Algorithms
Identifier: DOI
10.26300/zsx9-v648