BBOB Benchmarks¶
The Black-Box Optimization Benchmarking (BBOB) collection provides a comprehensive set of continuous optimization problems. Visit the COCO platform for a complete list of functions.
Available Suites¶
bbob:
bocode.BBOB.BBOBbbob-biobj:
bocode.BBOB.BBOB_Biobjbbob-biobj-mixint:
bocode.BBOB.BBOB_BiobjMixIntbbob-boxed:
bocode.BBOB.BBOB_Boxedbbob-constrained:
bocode.BBOB.BBOB_Constrainedbbob-largescale:
bocode.BBOB.BBOB_LargeScalebbob-mixint:
bocode.BBOB.BBOB_MixIntbbob-noisy:
bocode.BBOB.BBOB_Noisy
Available Functions¶
Sphere Function
Rastrigin Function
Schwefel Function
Griewank Function
Example Usage¶
import bocode
import torch
# Create a BBOB benchmark problem
problem = bocode.BBOB.BBOB(dim=5, function_number=2, instance_number=1) # Separable 5-dimensional ellipsoidal function
# Get problem information
bounds = problem.bounds
# Evaluate at a point
x = torch.Tensor([[0.0] * 5])
values, constraints = problem.evaluate(x)
print(f"BBOB function value at origin: {values[0]}")
Output:
BBOB function value at origin: tensor([42420381.6772])