Protection planning for critical infrastructure systems in location analysis: models and algorithms

Synopsis Supply systems are complex infrastructures devoted to the distribution of goods and services, composed by a combination of manufacturing, storage and transportation facilities. Because of their complexity, supply systems are highly vulnerable to a variety of threats. For example, they can be victim to natural disasters, the target of intentional disruptions or subject to accidental failures. In recent years a growing interest has been expressed in this eld, both in the academic and industrial communities. Some preliminary work showed that research has been mostly focusing on three main streams of investigation. One line of research has focused on the study of system design, which entails a full design of the networks so that they are intrinsically robust to disruptions. A dierent line of research has dealt with security issues, focusing on the identication of critical components through the use of interdiction models. The identication of critical system components is only the rst step, however, towards the development of sound and economically ecient protection strategies for systems which already exist and cannot be recongured. The third stream, thus, deals with protection of existing infrastructure systems. The main objective of this thesis is to contribute to the third stream of research, focused on protection modeling. This thesis provides a comprehensive review and a categorization of previous modeling approaches. The protection models proposed in the past are based on assumptions that strongly limit their applicability and/or identify strategies which are sub-optimal with respect to the complexity of real in-i frastructure systems. The main contribution of this thesis is to expand the protection literature by studying and solving new models that incorporate some realistic and complex features including uncertainty on the number of potential losses that a system may incur, capacity at the facility level, and correlated disruptions. Tailored solution approaches have been devised to solve the resulting complex mathematical models. This includes exact methodologies based on the generations of bounds, decomposition algorithms, and dualization and linearization procedures for the solution of multi-level optimization problems. An empirical analysis of the results has been run to demonstrate the usefulness of the models proposed in this thesis and the robustness of the solutions found. ii Acknowledgments I would like to express my gratitude to all those who gave me the possibility to work on this research project and that gave me their support. I am deeply indebted to my supervisor Dr. Maria Paola Scaparra whose …