Automatic Abstraction in Reinforcement Learning Using Ant System Algorithm.
Mohsen GhafoorianNasrin TaghizadehHamid BeigyPublished in: AAAI Spring Symposium: Lifelong Machine Learning (2013)
Keyphrases
- dynamic programming
- improved algorithm
- learning algorithm
- reinforcement learning
- experimental evaluation
- model free
- detection algorithm
- optimal solution
- computational cost
- linear programming
- high accuracy
- np hard
- k means
- search space
- objective function
- genetic algorithm
- cost function
- convergence rate
- matching algorithm
- ant colony optimization
- segmentation algorithm
- preprocessing
- optimization algorithm
- computationally efficient
- clustering method
- particle swarm optimization
- times faster
- machine learning
- function approximation
- feature selection
- evolutionary algorithm