Reinforcement Learning for Automatic Online Algorithm Selection - an Empirical Study.
Hans DegrooteBernd BischlLars KotthoffPatrick De CausmaeckerPublished in: ITAT (2016)
Keyphrases
- learning algorithm
- detection algorithm
- improved algorithm
- reinforcement learning
- times faster
- dynamic programming
- worst case
- high accuracy
- experimental evaluation
- np hard
- cost function
- computational cost
- multi robot
- selection algorithm
- significant improvement
- similarity measure
- function approximation
- recognition algorithm
- objective function
- monte carlo
- segmentation algorithm
- computational complexity
- optimal solution
- ant colony optimization
- expectation maximization
- online learning
- input data
- machine learning
- probabilistic model
- preprocessing
- policy search