Adaptive Reinforcement Learning: Increasing the applicability for large and time varying systems using parallel Gaussian Process regression and adaptive nonlinear control

This thesis investigates the applicability of the Probabilistic Inference for Learning COntrol (PILCO) algorithm to large systems and systems with time varying measurement noise. PILCO is a state-of-the-art model-learning Reinforcement Learning (RL) algorithm that uses a Gaussian Process (GP) model to average over uncertainties during learning....