Masters Theses
Abstract
"Resiliency is an important characteristic of any system. It signifies the ability of a system to survive and recover from unprecedented disruptions. Various characteristics exist that indicate the level of resiliency in a system. One of these attributes is the adaptability of the system. This adaptability can be enhanced by redundancy present within the system. In the context of system design, redundancy can be achieved by having a diverse set of good designs for that particular system. Evolutionary algorithms are widely used in creating designs for engineering systems, as they perform well on discontinuous and/or high dimensional problems. One method to control the diversity of solutions within an evolutionary algorithm is the use of combinatorial graphs, or graph based evolutionary algorithms. This diversity of solutions is key factor to enhance the redundancy of a system design. In this work, the way how graph based evolutionary algorithms generate diverse solutions is investigated by examining the influence of representation and mutation. This allows for greater understanding of the exploratory nature of each representation and how they can control the number of solution generated within a trial. The results of this research are then applied to the Travelling [sic] Salesman Problem, a known NP hard problem often used as a surrogate for logistic or network design problems. When the redundancy in system design is improved, adaptability can be achieved by placing an agent to initiate a transfer to other good solutions in the event of a disruption in network connectivity, making it possible to improve the resiliency of the system"--Abstract, page iii.
Advisor(s)
Corns, Steven
Committee Member(s)
Long, Suzanna, 1961-
Grasman, Scott E. (Scott Erwin)
Department(s)
Engineering Management and Systems Engineering
Degree Name
M.S. in Systems Engineering
Publisher
Missouri University of Science and Technology
Publication Date
Summer 2010
Pagination
ix, 63 pages
Rights
© 2010 Jayakanth Jayachandran, All rights reserved.
Document Type
Thesis - Open Access
File Type
text
Language
English
Subject Headings
Combinatorial enumeration problemsComputer networksEvolutionary computationGraph theory
Thesis Number
T 9671
Print OCLC #
688640533
Electronic OCLC #
648759802
Recommended Citation
Jayachandran, Jayakanth, "Improving resiliency using graph based evolutionary algorithms" (2010). Masters Theses. 4797.
https://scholarsmine.mst.edu/masters_theses/4797