Offline Reinforcement Learning for Optimizing Production Bidding Policies.
Dmytro KorenkevychFrank ChengArtsiom BalakirAlex NikulkovLingnan GaoZhihao CenZuobing XuZheqing ZhuPublished in: KDD (2024)
Keyphrases
- reinforcement learning
- optimal policy
- policy search
- control policies
- markov decision process
- markov decision processes
- optimal production
- reward function
- reinforcement learning agents
- function approximation
- partially observable markov decision processes
- state space
- reinforcement learning algorithms
- decision problems
- control policy
- fitted q iteration
- hierarchical reinforcement learning
- production system
- dynamic programming
- machine learning
- production planning
- multi agent
- real time
- model free
- partially observable
- inventory level
- learning algorithm
- decentralized control
- online auctions
- transfer learning
- infinite horizon
- sufficient conditions
- learning process
- continuous state
- bidding strategies
- policy gradient methods
- combinatorial auctions
- action space
- finite horizon
- production process
- average reward
- lost sales
- raw material
- optimal control
- temporal difference