Benchmarking genetic programming in a multi-action reinforcement learning locomotion task.
Ryan AmaralAlexandru IantaCaleidgh BayerRobert J. SmithMalcolm I. HeywoodPublished in: GECCO Companion (2022)
Keyphrases
- genetic programming
- reinforcement learning
- action selection
- symbolic regression
- fitness function
- evolutionary computation
- robot control
- gene expression programming
- reward shaping
- evolutionary algorithm
- partially observable domains
- action space
- function approximation
- machine learning
- financial forecasting
- transition model
- regression problems
- reinforcement learning algorithms
- markov decision processes
- continuous state
- multi agent
- genetic algorithm
- grammar guided genetic programming
- fitted q iteration
- degrees of freedom
- state space
- mobile robot
- learning algorithm
- rough terrain
- neural network