Detecting parameter change in dynamical systems using statistical methods

In this paper, different statistical methods for detecting parameter change in dynamical systems have been analysed. First, using the equations describing the system, different cases of errors have been identified. Using a t-test on the mean and a chi square test on the variance, we could detect these errors with a certain probability. The power on both of these tests is calculated to better these results. Furthermore, the time it takes for the system to reach a new steady state is analysed as well. Finally, the report gives an overview on the probabilities of detecting the right error in a dynamical system using statistical tests.