Interactive Playground
Configure parameters and compare optimization strategies in real-time
Select which benchmark mode you want to explore
Which surrogate model generated the benchmark landscape (GP is the standard setting; RF, NN, and BNN test robustness to different landscape structures)
Choose multiple optimizers to see how they perform side-by-side
How many experiments to run in parallel
Fraction of data points hidden from optimizers
Difficulty level of the optimization task
Which dataset to use for comparison
How to use: Results update automatically as you change any parameter — no button press needed. Optimizers are grouped by methodology (Bayesian GP → ANN → DOE → Direct Search) so the surrogate-model advantage is immediately visible. The Benchmark Configuration panel below the chart labels every setting used, making results directly comparable to figures in the paper. A few combinations are out of scope (neural-network surrogates were not run at the 99% hidden fraction) and will show no data.
Hide-the-Label — Results
Interactive visualization of benchmark results