A Comparison of Supervised and Reinforcement Learning Methods on a Reinforcement Learning Task

The \forward modeling" approach of Jor-dan and Rumelhart has been shown to be applicable when supervised learning methods are to be used for solving reinforcement learning tasks. Because such tasks are natural candidates for the application of reinforcement learning methods, there is a need to evaluate the relative merits of these two learning methods on reinforcement learning tasks. We present one such comparison here on a task involving learning to control an unstable, non-minimum phase, dynamic system. The comparison shows that the reinforcement learning method used performs better than the supervised learning method. An examination of the learning behavior of the two methods indicates that the diier-ences in performance can be attributed to the underlying mechanics of the two learning methods, which provides grounds for believing that similar performance diierences can be expected on other reinforcement learning tasks as well. This suggests that there is a set of tasks for which reinforcement learning methods are naturally applicable and more appropriate to use.