An Info-Gap Approach to Policy Selection for Bio-terror Response

Abstract

Bio-terror events are accompanied by severe uncertainty: great disparity between the best available data and models, and the actual course of events. We model this uncertainty with non-probabilistic information-gap models of uncertainty. This paper focuses on info-gaps in epidemiological models, in particular, info-gaps in the rate of infection. robustness to uncertainty is defined as a function of the required critical morbidity resulting from the attack. We show how preferences among available interventions are deduced from the robustness function. We demonstrate the irrevocable trade-off between robustness and demanded performance, and show that best-estimated performance has zero robustness. Finally, we present a theorem concerning the reversal of preferences between available interventions, and illustrate it with a numerical example.