Past, Present, and Future: An Optimal Online Algorithm for Single-Player GDL-II Games.
Florian GeißerThomas KellerRobert MattmüllerPublished in: ECAI (2014)
Keyphrases
- worst case
- dynamic programming
- optimal solution
- globally optimal
- cost function
- learning algorithm
- monte carlo
- preprocessing
- online learning
- computational cost
- optimal parameters
- k means
- computational complexity
- objective function
- np hard
- neural network
- lower bound
- detection algorithm
- matching algorithm
- markov decision processes
- closed form
- video games
- optimal strategy
- reinforcement learning
- passive aggressive