Year: 2023
Type: article
Abstract: The problem of finding a global minimum cut in an undirected graph is fundamental to combinatorial optimization.It has numerous applications including network reliability, clustering, and the Travelling Salesman Problem.In addition to this computational problem, structural and enumerative aspects of minimum cuts are also foundational to representation and algorithmic results.The number of constant-approximate global minimum cuts in a connected graph is polynomial in the number of vertices.This structural result has applications in constructing cut sparsifiers, sketching and streaming algorithms, and approximation algorithms for TSP.In this thesis we consider various generalizations of the minimum cut problem.We focus on solving these variants and on counting and enumerating optimum solutions.Our results include:• A new and faster algorithm for computing connectivity in hypergraphs,• The first deterministic polynomial time algorithm for enumerating hypergraph min-kcut-sets,• The first polynomial bound on the number of Multiobjective Min-cuts for a constant number of cost functions as well as the first polynomial time algorithm for enumerating all of them, and• Inapproximability results for representing symmetric submodular functions using hypergraph cut functions.Firstly, I would like to thank my advisor, Karthik Chandrasekaran.I have been inspired by Karthik's diligence and enthusiasm for research.I have benefited from his willingness to go through many iterations of revision to make every paper and talk (and particularly this thesis) the best it can possibly be.Karthik also introduced me to the fascinating area of minimum cuts and partitions which became the focus of this thesis.I would like to thank Chandra Chekuri, Sariel Har-Peled, and Sagnik Mukhopadhyay for serving on my thesis committee and for their valuable help and suggestions about this work.I am grateful to Chandra and Sagnik for their contributions to two of the papers on which this thesis is based, and to Sariel for engaging discussions on extensions of the work in this thesis which have led to further research.I also want to thank the others I have collaborated with in research, particularly Chao Xu, Weihang Wang, and Danupon Nanongkai, whose contributions to the work presented in this thesis were invaluable.I also want to thank to all the professors I have had the chance to serve as a teaching assistant for.I have learned a lot from the teaching styles and methods of the professors at Illinois.Special thanks to Brad Solomon for co-teaching CS 173 with me in the summer, and for giving me enough support to make the experience much less stressful, while treating me as an equal in running the course.I am very grateful for the entire theory group.I benefited both from discussions about our research, and from time spent sharing food or playing games.The group was a fun and welcoming place, and the camaraderie of my fellow students supported me through the challenges of grad school.Special thanks to Ian Ludden and Meghan Kelley for keeping me company in our office and to Qizheng He for our interesting discussions on min-cut in disk hypergraphs.I would never have pursued a PhD without the influence of faculty from Carnegie Mellon.Special thanks to Anil Ada, Ryan O'Donnell, and Wes Pegden who taught courses that sparked my love for theory and graphs, and who each mentored me in research projects which prepared me for my doctoral studies.Anil also mentored me over my six semesters as a TA for 15-251 and his passion for teaching was a great inspiration to me to become a professor.I also want to Jeremiah Blocki, who gave me my very first taste of research back in high school, and who motivated me to study computer science at CMU.Finally, I want to thank my parents, who raised me to love learning, and whose unfailing love and support carried me through the stress of my studies.
Author: Beideman, Calvin
Institution: University of Illinois Urbana-Champaign
Language: English
Subfield: Computational Theory and Mathematics
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7220769080
