Simulation Models for Cascading Effects in Critical Infrastructures

Contemporary critical infrastructures exhibit a strong interdependency, with many known direct implications of impacts (such as by blackouts), but also with a considerable number of indirect implications and side-effects to consider in the overall picture. Simulations of incidents and their propagation over subordinate and related structures of society is thus an important aid to understand the possibilities of damage and countermeasures, for an effective risk management. Suitable models differ in their expressiveness, flexibility, usability and domain specificity. While the energy sector allows for quite exact descriptions of effects using known physical dynamics, other structures like healthcare or social services are much less open to a precise mathematical description. These domains, particularly in combination with each other when they form a network of critical infrastructures, call for a higher level of abstraction, which induces tradeoffs between the aforementioned aspects.

This talk introduces two such models that arose from national and international research projects on critical infrastructure protection. Both are stochastic, and exhibit different properties regarding the dynamics that they can capture. The talk frames them into the landscape of general models for incident simulation, discusses pros and cons of different approaches, and concludes with the identification of open problems and proposed directions of research in the context of critical infrastructure models

Contemporary critical infrastructures exhibit a strong interdependency, with many known direct implications of impacts (such as by blackouts), but also with a considerable number of indirect implications and side-effects to consider in the overall picture. Simulations of incidents and their propagation over subordinate and related structures of society is thus an important aid to understand the possibilities of damage and countermeasures, for an effective risk management. Suitable models differ in their expressiveness, flexibility, usability and domain specificity. While the energy sector allows for quite exact descriptions of effects using known physical dynamics, other structures like healthcare or social services are much less open to a precise mathematical description. These domains, particularly in combination with each other when they form a network of critical infrastructures, call for a higher level of abstraction, which induces tradeoffs between the aforementioned aspects.

This talk introduces two such models that arose from national and international research projects on critical infrastructure protection. Both are stochastic, and exhibit different properties regarding the dynamics that they can capture. The talk frames them into the landscape of general models for incident simulation, discusses pros and cons of different approaches, and concludes with the identification of open problems and proposed directions of research in the context of critical infrastructure models

Medverkande

Upplysningar

  • Monika Magnusson
    Telefon: 054-700 2170
    E-post: monika.margnusson@kau.se