Henrik Hult: Rare events, large deviations, and stochastic simulation

Quantifying probabilities of rare events is a central issue of risk assessment. I will present recent progress on design and analysis of efficient stochastic simulation algorithms for computing probabilities of rare events. The analysis builds on connections between stochastic control theory, large deviations and Hamilton-Jacobi equations. I will present some applications to chemistry, finance, and power systems. ​

Remark: This is the first lecture in a series of "internal colloquia" at KTH. The intended audience is all members in the department (in particular, members from other divisions who would like find out what their colleagues are working on are most welcome.)