List your company

Operational Throughput Optimization on Internal Learned Simulators (held-out shifts)

2.4Throughput Improvement (%)Offline Actor-Critic
-0.0960.5521.21.848Mar 25, 2026

Evaluation results

MethodLinks
Offline Actor-Critic2026.032.42.3
Fine-Tuned Behavior Cloning2026.032.12
Behavior Cloning2026.031.61.4
Human Decisions (replayed)2026.030—