On Thursday October 4, 2018, at 10:00 in room 693, Prof. Boris Houska from ShanghaiTech University, China, will give a scientific talk on " Gram-Charlier Expansion Methods and their Applications in Control".
This talk is about stochastic linear discrete-time processes.
We assume that the process noise sequence consists of independent random
variables that are identically and uniformly distributed on given compact
zonotopes. Because the central limit theorem fails to hold for the
associated probability distributions of the state sequence, we propose a
computationally tractable method for approximating both the transient and
limit distribution of the state sequence by using a generalized Gram-Charlier
series with respect to a multivariate Chebyshev basis. The corresponding
approach finds applications in the field of robust linear control design
for stochastic systems.