Optimal cardiac pacing with Q learning

A cardiac pacemaker control system constituted of: a means for receiving input from a hemodynamic sensor; an adaptive control system in communication with the means for receiving input from the hemodynamic sensor; and an interface arranged to provide cardiac stimulation responsive to the adaptive control system; the adaptive control system comprising a learning module operative to converge to patient specific cardiac pacing stimulation timing using a machine learning scheme in cooperation with a probabilistic replacement scheme, the probabilistic replacement scheme arranged to replace inputs from the hemodynamic sensor with online calculated values.