In this article, we explore an evolutionary approach to the optimization of potential-based shaping rewards and meta-parameters in reinforcement learning. Shaping rewards is a frequently used approach to increase the learning performance of reinforcement learning, with regards to both initial performance and convergence speed. Shaping rewards provide additional knowledge to the agent in the form of richer reward signals, which guide learning to high-rewarding states. Reinforcement learning depends critically on a few meta-parameters that modulate the learning updates or the exploration of the environment, such as the learning rate α, the discount factor of future rewards γ, and the temperature τ that controls the trade-off between exploration and exploitation in softmax action selection. We validate the proposed approach in simulation using the mountain-car task. We also transfer shaping rewards and meta-parameters, evolutionarily obtained in simulation, to hardware, using a robotic foraging task.
Ackley, D.H., & Littman, M.L. (1991). Interactions between learning and evolution. In C. G. Langton, C. Taylor, C. D. Farmer, & S. Rasmussen (Eds.), Artificial Life II: Santa Fe Institute Studies in the Sciences of Complexity (Vol. 10, pp. 487-509). Redwood City, CA: Addison-Wesley.
Doya, K., & Uchibe, E. (2005). The cyber rodent project: Exploration of adaptive mechanisms for self-preservation and self-reproduction. Adaptive Behavior, 13(2), 149-160.
4.
Elfwing, S., Uchibe, U., Doya, K., & Christensen, H.I. (2005). Biologically inspired embodied evolution of survival. In Proceedings of the IEEE Congress on Evolutionary Computation, CEC2005 (Vol. 3, pp. 2210-2216). Piscataway, NJ: IEEE.
5.
Elfwing, S., Uchibe, U., Doya, K., & Christensen, H.I. (2007a). Darwinian embodied evolution of the learning ability for survival. Manuscript submitted for publication.
6.
Elfwing, S., Uchibe, U., Doya, K., & Christensen, H.I. (2007b). Evolutionary development of hierarchical learning structures. IEEE Transactions on Evolutionary Computation, 11(2), 249-264.
7.
Eriksson, A., Capi, G., & Doya, K. (2003). Evolution of meta-parameters in reinforcement learning algorithm. In Proceedings of the IEEE/RSJ Conference on Intelligent Robots and Systems, IROS2003 (pp. 412-417). Piscataway, NJ: IEEE.
8.
Konidaris, G., & Barto, A. (2006). Autonomous shaping: Knowledge transfer in reinforcement learning. In Proceedings of the International Conference on Machine Learning, ICML2006 (pp. 489-496). New York: ACM.
9.
Laud, A., & DeJong, G. (2002). Reinforcement learning and shaping: Encouraging intended behaviors. In Proceedings of the International Conference on Machine Learning, ICML2002 (pp. 355-362). San Francisco, CA: Morgan Kaufmann.
10.
Laud, A., & DeJong, G. (2003). The influence of reward on the speed of reinforcement learning: An analysis of shaping. In Proceedings of the International Conference on Machine Learning, ICML2003 (pp. 440-447). San Francisco, CA: Morgan Kaufmann.
11.
Marthi, B. (2007). Automatic shaping and decomposition of reward functions. In Proceedings of the International Conference on Machine Learning, ICML2007 (pp. 601- 608). New York: ACM.
12.
Mataric, M. (1994). Reward functions for accelerated learning. In Proceedings of the International Conference on Machine Learning, ICML1994 (pp. 181-189). New York: ACM.
13.
Mataric, M. (1997). Reinforcement learning in the multirobot domain . Autonomous Robots, 4(1), 73-83.
14.
Moriarty, D.E., Schultz, A.C., & Grefenstette, J.J. (1999). Evolutionary algorithms for reinforcement learning. Journal of Artificial Intelligence Research, 11, 241-276.
15.
Ng, A.Y., Harada, D., & Russell, S.J. (1999). Policy invariance under reward transformations: Theory and application to reward shaping. In Proceedings of the International Conference on Machine Learning, ICML1999 (pp. 278-287). San Francisco, CA: Morgan Kaufmann.
16.
Nolfi, S., & Floreano, D. (2000). Evolutionary robotics. The biology, intelligence, and technology of self-organizing machines. Cambridge, MA: MIT Press.
17.
Randløv, J., & Alstrøm, P. (1998). Learning to drive a bicycle using reinforcement learning and shaping. In Proceedings of the International Conference on Machine Learning, ICML1998. San Francisco, CA: Morgan Kaufmann.
18.
Rummery, G.A., & Niranjan, M. (1994). On-line Q-learning using connectionist systems. Technical Report CUED/F-INFENG/ TR 166, Cambridge University Engineering Department.
Skinner, B.F. (1938). The behavior of organisms: An experimental analysis . Englewood Cliffs, NJ: Prentice-Hall .
21.
Stanley, K.O., & Miikkulainen, R. (2002). Effcient reinforcement learning through evolving neural network topologies. In Proceedings of the Genetic and Evolutionary Computation Conference, GECCO2002 (pp. 569-577). San Francisco, CA: Morgan Kaufmann.
22.
Sutton, R.S. (1996). Generalization in reinforcement learning: Successful examples using sparse coarse coding. In D. S. Touretzky, M. C. Mozer, & M. E. Hasselmo (Eds.), Advances in Neural Information Processing Systems 8 (pp. 1038-1044). Cambridge, MA: MIT Press.
23.
Sutton, R.S., & Barto, A. (1998). Reinforcement learning: An introduction. Cambridge, MA: MIT Press.
24.
Watson, R., Ficici, S., & Pollack, J. (2002). Embodied evolution: Distributing an evolutionary algorithm in a population of robots. Robotics and Autonomous Systems, 39(1), 1-18.
25.
Whiteson, S., & Stone, P. (2006). Evolutionary function approximation for reinforcement learning. Journal of Machine Learning Research, 7, 877-917.
26.
Wiewiora, E. (2003). Potential-based shaping and Q-value initialization are equivalent. Journal of Artificial Intelligence Research , 19, 205-208.
