Strategies of control in a linear stochastic system with non-Gaussian disturbances (Q1913852)

Consider a discrete-time linear stochastic system with non-Gaussian disturbances where observations are given. The observations are linear transformations of the states disturbed by non-Gaussian noise too. The admissible controls are feedback strategies \(\psi n(\cdot)\) of the observations. The quadratic quality criterion is interpreted as a function of the controls and the distribution of the initial condition and noise terms. Optimal and suboptimal minimax strategies are determined.