Reward Function and Initial Values: Better Choices for Accelerated Goal-Directed Reinforcement Learning
An important issue in Reinforcement Learning (RL) is to accelerate or improve the learning process. In this paper, we study the influence of some RL parameters over the learning speed. Indeed, although RL convergence properties have been widely studied, no precise rules exist to correctly choose the reward function and initial Q-values. Our method helps the choice of these RL parameters within the context of reaching a goal in a minimal time. We develop a theoretical study and also provide experimental justifications for choosing on the one hand the reward function, and on the other hand particular initial Q-values based on a goal bias function.
KeywordsOptimal Policy Reinforcement Learning Goal State Reward Function Systematic Exploration
Unable to display preview. Download preview PDF.
- 1.Sutton, R.S., Barto, A.G.: Reinforcement Learning: An Introduction. The MIT Press, Cambridge (1998)Google Scholar
- 2.Watkins, C.: Learning from Delayed Rewards. PhD thesis, Cambridge University, Cambridge, England (1989)Google Scholar
- 3.Mataric, M.J.: Reward functions for accelerated learning. In: Proc. of the 11th ICML, pp. 181–189 (1994)Google Scholar
- 4.Ng, A.Y., Harada, D., Russell, S.: Policy invariance under reward transformations: theory and application to reward shaping. In: Proc. of the 16th ICML, pp. 278–287 (1999)Google Scholar
- 6.Hailu, G., Sommer, G.: On amount and quality of bias in reinforcement learning. In: Proc. of the IEEE International Conference on Systems, Man and Cybernetics, Tokyo, pp. 1491–1495 (1999)Google Scholar
- 8.Behnke, S., Bennewitz, M.: Learning to play soccer using imitative reinforcement. In: Proc. of the ICRA Workshop on Social Aspects of Robot Programming through Demonstration, Barcelona (2005)Google Scholar