PhD thesis 'Extreme value statistics and optimization problems in stochastic processes'

This thesis is devoted to the study of extreme value statistics in stochastic processes and their applications. In the first part, we obtain exact analytical results on the extreme value statistics of both discrete-time and continuous-time random walks. In particular, we focus on the gap statistics of random walks and exhibit their asymptotic universality with respect to the jump distribution in the limit of a large number of steps. In addition, we compute the asymptotic behavior of the expected maximum of random walks in the presence of a bridge constraint and reveal a rich behavior in their finite-size correction. Moreover, we compute the expected length of the convex hull of Brownian motion confined in a disk and show that it converges slowly to the perimeter of the disk with a stretched exponential decay. In the second part, we focus on numerically sampling rare trajectories of stochastic processes. We introduce an efficient method to sample bridge discrete-time random walks. We illustrate it and apply it to various examples. We further extend the method to other stochastic processes, both Markovian and non-Markovian. We apply our method to sample surviving particles in the presence of a periodic trapping environment. Finally, we discuss several optimization problems in stochastic processes involving extreme value statistics. In particular, we introduce a new technique to optimally control dynamical systems undergoing a resetting policy.
Comment: 134 pages, 54 figures