Scalable High-Dimensional Reachable Set Estimation

Comparing estimation algorithms under the curse of dimensionality for gamma-quantized reachability

0.841
Best F1 Score (d=2)
10^15
PAC Bound at d=100
4
Algorithms Compared
2-100
Dimensions Tested

F1 Score vs. Dimensionality

F1 Score vs. Sample Budget

Sample Complexity Growth

Algorithm Summary

Detailed Results

AlgorithmMean F1Best F1Worst F1Scalability
MC PAC0.2630.8410.000Poor
Adaptive MC0.2510.8130.000Moderate
DimRed + MC0.1590.8410.000Good
Learned Boundary0.1380.8270.000Good