Use this URL to cite or link to this record in EThOS: https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.800206
Title: Graph partitions with proportional density and colouring constraints
Author: Dallard, Clément
ISNI:       0000 0004 8507 9925
Awarding Body: University of Portsmouth
Current Institution: University of Portsmouth
Date of Award: 2019
Availability of Full Text:
Access from EThOS:
Access from Institution:
Abstract:
Motivated by real life challenges and pure research curiosity, this thesis considers several combinatorial problems related to combinatorial optimisation, computational complexity and graph theory. In each chapter, our goal is to refine our understanding of the complexity of the studied problem. When considering N Phardproblems, we eventually circumvent the expected exponential complexity by providing parameterized and approximation algorithms. We also study the problems on restricted type of instances and propose efficient polynomial-time algorithms. We study the notions of proportional density and define a proportionally dense subgraph as a subgraph whose vertices have proportionally as many neighbours inside the subgraph as in the whole graph. This notion combines local and global properties of the subgraph, an interesting paradigm rarely encountered in graph theory. Two problems related to proportionally dense subgraphs are studied, from the perspectives of structural graph theory, computational complexity and approximation. Then, we focus on a graph partitioning problem on vertex-coloured graphs, and study its complexity on restricted classes of caterpillars with bounded hair-length and planar graphs. Our contributions expose a gap in the complexity of the problem with regard to the hair-length and the maximum degree of the graph. Lastly, we propose a complexity study of a scheduling problem with fixed route and soft time constraints. We show that the problem’s complexity inherently depends on the ride time constraints and give several polynomial-time algorithms for restricted type of instances.
Supervisor: Bader-El-Den, Mohamed ; Chlebikova, Janka ; Gegov, Alexander Emilov Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.800206  DOI: Not available
Share: