Deep Reinforcement Learning for Crowdsourced Urban Delivery: System States Characterization, Heuristics-guided Action Choice, and Rule-Interposing Integration.
Tanvir AhamedBo ZouNahid Parvez FaraziTheja TulabandhulaPublished in: CoRR (2020)
Keyphrases
- reinforcement learning
- state action
- initial state
- transition model
- perceptual aliasing
- action selection
- action sequences
- action rules
- state space
- partially observable domains
- state abstraction
- condition action rules
- optimal policy
- markov decision problems
- multi agent
- state transitions
- heuristic function
- action space
- derived predicates
- markov decision process
- model free
- state transition
- rule sets
- heuristic search
- association rules
- search algorithm
- reinforcement learning algorithms
- reward shaping
- belief state
- function approximation
- human actions
- hyper heuristics
- classification rules
- markov decision processes
- partial knowledge
- machine learning
- reward signal
- video annotation
- rule discovery
- learning classifier systems
- low quality
- learning process
- learning algorithm