POMDP Abstraction on Tiger full actions
0Validation Error Rateε-quotient
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| ε-quotientepsilon=0.3, m=1, T=22026.03 | 0 | 0 | — | 4 | |
| Truncation (d=1)epsilon=0.3, m=1, T=22026.03 | 0 | 0 | 0.01 | 5 | |
| Belief-distanceepsilon=0.3, m=1, T=22026.03 | 0 | 0 | 0.01 | 6 | |
| Bisimulationepsilon=0.3, m=1, T=22026.03 | 0 | 0 | 0.05 | 6 | |
| Truncation (d=1)epsilon=0.5, m=1, T=22026.03 | 0 | 0 | 0.01 | 5 | |
| Belief-distanceepsilon=0.5, m=1, T=22026.03 | 0 | 0 | 0.01 | 5 | |
| Bisimulationepsilon=0.5, m=1, T=22026.03 | 0 | 0 | 0.05 | 5 | |
| Random partitionepsilon=0.3, m=1, T=22026.03 | 4.1 | 0.082 | 0.01 | 3.7 | |
| ε-quotientepsilon=0.5, m=1, T=22026.03 | 12.2 | 0.245 | — | 3 | |
| Random partitionepsilon=0.5, m=1, T=22026.03 | 12.2 | 0.245 | 0.01 | 3 |