RLWS: A Reinforcement Learning based GPU Warp Scheduler.
Jayvant AnantpurNagendra Dwarakanath GulurShivaram KalyanakrishnanShalabh BhatnagarR. GovindarajanPublished in: CoRR (2017)
Keyphrases
- reinforcement learning
- function approximation
- real time
- model free
- reinforcement learning algorithms
- graphics hardware
- temporal difference
- temporal difference learning
- parallel implementation
- multi agent
- state space
- parallel computation
- scheduling algorithm
- learning algorithm
- graphics processing units
- parallel computing
- gpu implementation
- markov decision processes
- learning process
- function approximators
- stochastic approximation
- heterogeneous computing
- multi agent reinforcement learning
- graphics processors
- real robot
- action selection
- optimal control
- parallel processing
- supervised learning
- genetic algorithm
- neural network