Addressing Reward Engineering for Deep Reinforcement Learning on Multi-stage Task.
Bin ChenJianhua SuPublished in: ICONIP (5) (2019)
Keyphrases
- multistage
- reinforcement learning
- optimal policy
- dynamic programming
- single stage
- stochastic programming
- function approximation
- engineering design
- lot sizing
- eligibility traces
- production system
- learning algorithm
- reinforcement learning algorithms
- state space
- model free
- reward function
- stochastic optimization
- attack detection
- machine learning
- reinforcement learning methods
- markov decision processes
- partially observable environments
- average reward
- multi agent
- temporal difference
- production line
- learning agent
- state action
- action selection
- finite state
- policy gradient
- decision problems
- search algorithm
- reward shaping