Trial Comparison
Parameters
Metrics
foamBO
Parent:
Specialized at:
Context:
Scenarios:
Fidelity:
Target:
Levels:
Cost metric:
Trials per fidelity:
Completed
Running
Failed
Early Stopped
Generation Strategy
●
Budget
Model Fitted
Execution Timing
Total:
Trial execution:
Generation:
Overhead:
Dimensionality Reduction
Monitored via generation strategy
Compute Efficiency
Avg Trial Duration
Parallelism
Event Log
| Trial | Status | Gen Node | Fidelity | Source | Case | Time | ||
|---|---|---|---|---|---|---|---|---|
|
Parameters
Log
Visualize
sys.argv[1] → case path, sys.argv[2] → screenshot filename (saved in case folder)
|
Hypervolume Trace
Robustness Profile
How stable is each Pareto-optimal design across environmental conditions. Lower spread = more robust.
Risk Measure Trace
CVaR per trial over time. Rising = finding more robust designs.
Context Sensitivity
Which context variables matter for which objectives (std-dev when sweeping each context param).
Convergence
Per-objective Probability of Improvement (PI). Lower PI = closer to converged. Threshold:
PI:
Specialization: up to additional trials
(worst at , PI=)
Already specialized at all contexts
Specialization cost per context scenario
| Context | |
|---|---|
Per-trial distance from the running best, normalized by IQR. 0 = matched or set the best. Values trending toward 0 indicate the optimizer is consistently near the optimum. Dashed line = improvement threshold from stopping strategy.
Context Scenarios
Sobol-sampled environmental conditions used for robust optimization. Parameters:
| # | |
|---|---|
No context points available.
Group Sensitivity
Group Interactions
Click "Generate Insights" to compute model-based analyses. Requires a fitted model.
Early Stopped:
No streaming data yet
No trial dependencies configured
Dependencies configured but none triggered yet (all skipped)
Pick Interesting Point
Select the best Pareto point for a given objective (best value + lowest uncertainty):
Parameter Input
Freeze:
Group "": best = T ( matches)
Predictions
Robust Analysis
Nominal context:
Nominal
CVaR
Gap
Spread
Per-scenario breakdown
| # | ||
|---|---|---|
Partial Dependence
🔒
=