Fuzzy Systems : agreement and modeling : with applications to Critical Infrastructure Protection problems.

Abstract

Protecting Critical Infrastructures is a hard task, since these systems are becoming more and more interdependent for many reasons, while the knowledge on their behavior is becoming more and more sector speci c. In order to provide instruments for their protection, a rst step is to de ne convincing frameworks able to assess and predict the evolution of their working condition by sharing information among infrastructures on a real-time basis. Moreover, due to the lack of adequate quantitative data about the interaction among infrastructures and their subsystems, there is the need to provide formalisms able to handle the information elicited by experts and stakeholders, which typically express in a linguistic and vague way. This was the aim of MICIE European project, where a distributed real-time framework based on a fuzzy interdependency model was adopted. This Thesis is aimed to summarize the e orts done within such a project, following three main directions. First of all, a formal framework for the analysis of systems characterized by vagueness and ambiguity is provided; some of the most studied distributed agreement problems are then addressed in the fuzzy fashion, thus providing a theoretical justi cation of the MICIE project; nally, some modeling methodologies, aimed to account for vagueness and fuzziness, are nally discussed, along with the online approach implemented within the MICIE project.