An efficient data-based off-policy Q-learning algorithm for optimal output feedback control of linear systems

We provide supplementary code material to our recent paper "An efficient data-based off-policy Q-learning algorithm for optimal output feedback control of linear systems".

The provided matlab codes contain a comparison between our proposed Qlearning algorithm and an existing SDP-based method. To reproduce the results of Table I, you can run the file 'comparison_QL_SDP.m'. Note that CVX is required to run the SDP-based method, see (http://cvxr.com/cvx/) for details.