Long-Horizon Control on Industrial Simulator Offline (test)
21.2ThroughputRandom
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| RandomDescription=no-control policy2026.06 | 21.2 | 12.3 | 10.8 | |
| DQLTraining Objective=Event-driven mean aggregation (Eq. 12)2026.06 | 18 | 16 | 12.8 | |
| CQL(Q*)Training Objective=Event-driven mean aggregation (Eq. 12), Variant=Q-learning variant2026.06 | 17.5 | 15.7 | 12.3 | |
| IQLTraining Objective=Event-driven mean aggregation (Eq. 12)2026.06 | 14.7 | 13 | 10.6 | |
| SPTDescription=dispatching heuristic (shortest processing time)2026.06 | 13.9 | 13.3 | 10.7 |