Fast Population-Based Reinforcement Learning on a Single Machine.
Arthur FlajoletClaire Bizon MonrocKarim BeguirThomas PierrotPublished in: ICML (2022)
Keyphrases
- single machine
- reinforcement learning
- scheduling problem
- dynamic programming
- total weighted tardiness
- minimize total
- processing times
- maximum lateness
- earliness tardiness
- release times
- release dates
- total tardiness
- scheduling jobs
- production scheduling
- optimal policy
- state space
- learning effect
- learning algorithm
- rolling horizon
- markov decision processes
- competitive ratio
- weighted tardiness
- weighted number of tardy jobs
- setup times
- sequence dependent setup times
- combinatorial optimization
- particle swarm optimization
- simulated annealing
- machine learning
- differential evolution
- np hard
- special case
- search space
- deteriorating jobs