Solution of the linear quadratic regulator problem of black box linear systems using reinforcement learning

In this paper, a Q-learning algorithm is proposed to solve the linear quadratic regulator problem of black box systems. The only has access input and output measurements. A Luenberger observer parametrization constructed using control new obtained from factorization utility function. An integral reinforcement learning approach used develop approximator structure. gradient descent update rule estimate on-line parameters Q-function. Stability convergence under assessed Lyapunov stability theory. Simulation studies are carried out verify approach.