Accelerating flat reinforcement learning on a robot by using subgoals in a hierarchical framework
B. Van Vliet, Wouter Caarls, Erik Schuitema, Pieter P. Jonker · Research Repository (Delft University of Technology) · 2010
Learning a motor skill task with Reinforcement Learning still takes a long time. A way to speed up the learning process with- out using much prior knowledge is to use sub-goals. In this study, the use of subgoals decreased the learning time by a factor nine and we show that tests on a real robot give similar results. The price to be paid, in case the subgoals do not lie on the optimal path, is a worse end performance. Hierarchical greedy execution can (partially) cancel out this problem. For future work, we suggest the use of a method which is able to obtain optimal performance.