Stabilization of a linear switched system with a data-driven approach

With the rising of automation in our daily lives, we observe more and more hybrid systems around us. Those systems can be too complex for us to model without adding approximation errors. A solution is to use the data-driven approach to directly control them. Our focus here will be on a subfamily of hybrid systems: linear switched systems. We will use the gradient descent method as well as the Nelder-Mead simplex method to address the problem of stabilizing such a system. An analysis on their performances is also provided.