Multi-Stage Temporal Difference Learning for 2048-like Games.
Kun-Hao YehI-Chen WuChu-Hsuan HsuehChia-Chuan ChangChao-Chin LiangHan ChiangPublished in: CoRR (2016)
Keyphrases
- multistage
- temporal difference learning
- game playing
- monte carlo tree search
- function approximation
- fixed point
- evaluation function
- video games
- reinforcement learning
- dynamic programming
- single stage
- temporal difference
- markov decision process
- lot sizing
- game play
- computer games
- reinforcement learning algorithms
- optimal policy
- machine learning
- np hard
- average cost
- action selection
- serious games
- least squares
- attitudes toward
- convergence rate
- learning outcomes