Continuous Black-Box Optimization (C-BBO) benchmarks for DeepHyper.
| Function Name | Number of Dimensions | Comment |
|---|---|---|
| ackley |
|
Many local minima and single global optimum |
| branin | 2 | Three global optimum |
| cossin | 1 | Many local minima, good for visualisation. |
| easom | 2 | Almost flat everywhere |
| griewank |
|
|
| hartmann6D | 6 | |
| levy |
|
|
| michal |
|
|
| rosen |
|
|
| schwefel |
|
|
| shekel | 4 | Many local minima with flat areas |
Python installation and dependency management is handled with uv. Clone this repository then create a Python environment with uv sync.
Go to the example directory and run the benchmarks with uv run benchmark cbbo.toml. Plot the results of the benchmarks with uv run benchmark cbbo.toml --plot.