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.BBOB

  • bbob-biobj: bocode.BBOB.BBOB_Biobj

  • bbob-biobj-mixint: bocode.BBOB.BBOB_BiobjMixInt

  • bbob-boxed: bocode.BBOB.BBOB_Boxed

  • bbob-constrained: bocode.BBOB.BBOB_Constrained

  • bbob-largescale: bocode.BBOB.BBOB_LargeScale

  • bbob-mixint: bocode.BBOB.BBOB_MixInt

  • bbob-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])