Point-to-Point Car Racing: an Initial Study of Evolution Versus Temporal Difference Learning
- 1 January 2007
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- No. 23254270,p. 260-267
Abstract
This paper considers variations on an extremely simple form of car racing, the challenge being to visit as many way-points as possible in a fixed amount of time. The simplicity of the models enables a very thorough evaluation of various learning algorithms and control architectures, and enables other researchers to work on the same models with relative ease. The models are used to compare the performance of various hand-programmed controllers, and neural networks trained using evolution, and using temporal difference learning. Comparisons are also made between state-based and action-based controller architectures. The best controllers were obtained using evolution to learn the weights of state-evaluation neural networks, and these were greatly superior to human driversKeywords
This publication has 10 references indexed in Scilit:
- Stanley: The robot that won the DARPA Grand ChallengeJournal of Field Robotics, 2006
- Evolving robust and specialized car racing skills2006
- Temporal Difference Learning Versus Co-Evolution for Acquiring Othello Position Evaluation2006
- Improving Artificial Intelligence In a Motocross Game2006
- Evolving Controllers for Simulated Car Racing2005
- Evolution of the Driving Styles of Anticipatory Agent Remotely Operating a Scaled Model of Racing Car2005
- Coevolution Versus Self-Play Temporal Difference Learning for Acquiring Position Evaluation in Small-Board GoIEEE Transactions on Evolutionary Computation, 2005
- Coevolution of active vision and feature selectionBiological Cybernetics, 2004
- Apprenticeship learning via inverse reinforcement learning2004
- Reinforcement Learning: An IntroductionIEEE Transactions on Neural Networks, 1998