Reinforcement learning and evolutionary algorithms for non-stationary multi-armed bandit problems.
Dimitris E. KoulouriotisA. S. XanthopoulosPublished in: Appl. Math. Comput. (2008)
Keyphrases
- non stationary
- evolutionary algorithm
- reinforcement learning
- multi armed bandit problems
- multi objective
- evolutionary computation
- optimization problems
- fitness function
- multi objective optimization
- differential evolution
- adaptive algorithms
- differential evolution algorithm
- bandit problems
- genetic algorithm
- white noise
- state space
- temporal evolution
- markov decision processes
- autoregressive
- stock price
- dynamic programming
- machine learning
- nsga ii
- empirical mode decomposition
- learning algorithm
- optimal policy
- special case
- lower bound
- change point detection