Enhancing Reinforcement Learning Performance in Delayed Reward System Using DQN and Heuristics.
Keecheon KimPublished in: IEEE Access (2022)
Keyphrases
- reinforcement learning
- function approximation
- eligibility traces
- state space
- model free
- control strategies
- reinforcement learning algorithms
- heuristic search
- average reward
- partially observable environments
- machine learning
- temporal difference
- learning algorithm
- multi agent
- search algorithm
- optimal control
- action selection
- total reward
- markov decision processes
- optimal policy
- supervised learning
- dynamic programming
- learning process
- control policy
- reinforcement learning methods
- partially observable
- robotic control
- multi armed bandit
- markov decision problems
- learning agent
- reward function
- evolutionary algorithm
- learning problems