Survivability and Reciprocal Altruism: Two Strategies for Intelligent Infrastructure with Applications to Power Grids

While electric power grids are generally robust to small failures and thus provide a fairly high level of reliability, they are notably vulnerable to large, spectacular cascading failures. Single component failures rarely impede the ability of a power grid to serve its customers. But larger sets of concurrent outages can produce blackouts with sizes that are highly improbable from the perspective of Gaussian statistics. Because of the number of components in a power grid it is impossible to plan for and mitigate all sets of failures. Maintaining a high level of reliability in the midst of this risk is challenging. As market forces, variable sources (e.g., wind and solar power) and new loads (e.g., electric cars) increase stress on electricity infrastructure, the challenge of managing grid reliability and costs will certainly increase. Therefore we need strategies that enable the most important services that depend on electricity infrastructure to continue in the midst of risks. This chapter discusses two strategies for enabling the most important services that depend on electricity continue in the midst of significant systemic vulnerability. The first, as proposed by Talukdar et al. [26] is survivability, in which backup electricity sources provide a very high level of reliability for services that are economically and socially vital. The second, as proposed by Hines et al. [16], is Reciprocal Altruism, under which agents that manage the infrastructure are encouraged to align personal goals with those of the system as a whole.