An iterative batch RL method combining model-based policy search with minimum pairwise trajectory diversity and behavior-based safety constraints speeds up cost reduction across batch iterations in Industrial Benchmark experiments.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Iterative Batch Reinforcement Learning via Safe Diversified Model-based Policy Search
An iterative batch RL method combining model-based policy search with minimum pairwise trajectory diversity and behavior-based safety constraints speeds up cost reduction across batch iterations in Industrial Benchmark experiments.