Modeling network vulnerability over space and time
Abstract
[ACCESS RESTRICTED TO THE UNIVERSITY OF MISSOURI AT REQUEST OF AUTHOR.] Transportation networks play an important role in moving freight, people, data and energy. Therefore, ensuring continuity in these services is an essential goal in network planning and management. However, due to limited resources, efforts to safeguarding system operation must be prioritized based on assessments of where vulnerabilities may exist. Many network vulnerability assessment approaches have been designed to characterize system vulnerability to disruption and to identify those components (arcs/nodes) that may contribute most to system vulnerability. However, as network complexity increases, it becomes increasingly difficult to account for all of the interrelationships supported by the network and to determine the contribution of network components to system vulnerability. Furthermore, nodal interactions supported by networks can vary over time and space, exacerbating the difficulty of identifying and characterizing potential vulnerabilities. In this thesis, a network optimization model is extended to aid in the search for scenarios of arc and/or node loss that represent the largest vulnerabilities in networked systems. To increase the model's applicability to large-scale systems, a constraint/variable reduction technique is proposed to further improve its computational characteristics. The developed modeling framework is applied to a large-scale Internet system. Results highlight the ability of the proposed framework to efficiently assess network vulnerability as well as the sensitivity of characterizations of network vulnerability to the dynamic nature of nodal flows over space and time.
Degree
M.S.
Thesis Department
Rights
Access to files is limited to the University of Missouri--Columbia.