On Finding Optimal Sub-structures in Graphs

Date of Submission

July 2022

Date of Award

7-1-2023

Institute Name (Publisher)

Indian Statistical Institute

Document Type

Doctoral Thesis

Degree Name

Doctor of Philosophy

Subject Name

Mathematics

Department

Advance Computing and Microelectronics Unit (ACMU-Kolkata)

Supervisor

Nandy, Subhas Chandra (ACMU-Kolkata; ISI)

Abstract (Summary of the Work)

In computer science, a problem is said to have an optimal sub-structure if an optimal solution can be constructed from optimal solutions of its sub-problems. These optimal sub-structures are computed in the classical graph-theoretic setting where the graph is a structure with a set of vertices and edges. In computational geometry, the vertex set is usually represented by a set of geometric objects like unit disks, etc., and the edge set is represented by the intersection of these geometric structures. In this thesis, three problems are investigated namely minimum discriminating codes, red-blue separation, and minimum consistent subset. In the minimum discriminating codes problem, we handle some geometric structures like unit intervals and arbitrary intervals in $\IR$ and axis parallel unit squares in $\IR^2$. We prove the hardness of the problem in both one-dimensional and two-dimensional planes. We also propose PTAS for the unit interval case and a 2-factor approximation algorithm for the arbitrary interval case. In polynomial time we have given approximation algorithms producing constant-factor solution in $\IR^2$ with axis parallel unit square objects. We have also studied a similar problem known as the minimum identifying codes in some geometric settings. In the red-blue separation problem, we consider a graph whose vertices are coloured red or blue. We study the computational complexity in some graph classes. We design polynomial-time algorithms when one of the coloured classes is bounded by a constant. We also give some tight bounds on the cardinality of the optimal solution. In the minimum consistent subset problem, we work with simple graph classes like paths, caterpillars, trees, etc. For each of these graphs, we have designed optimal algorithms. We have also considered both undirected and directed versions for a few of the graphs

Comments

ProQuest Collection ID: https://www.proquest.com/pqdtlocal1010185/dissertations/fromDatabasesLayer?accountid=27563

Control Number

ISILib-TH549

Creative Commons License

Creative Commons Attribution 4.0 International License
This work is licensed under a Creative Commons Attribution 4.0 International License.

DOI

http://dspace.isical.ac.in:8080/jspui/handle/10263/2146

Link to Full Text

Share

COinS