Himmelblau Benchmark

Imports

import os

import pandas as pd

import bofire.strategies.api as strategies
from bofire.benchmarks.single import Himmelblau
from bofire.data_models.acquisition_functions.api import qLogEI
from bofire.data_models.api import Domain
from bofire.data_models.strategies.api import RandomStrategy, SoboStrategy
from bofire.runners.api import run


SMOKE_TEST = os.environ.get("SMOKE_TEST")

Random Optimization

def sample(domain):
    datamodel = RandomStrategy(domain=domain)
    sampler = strategies.map(data_model=datamodel)
    sampled = sampler.ask(10)
    return sampled


def best(domain: Domain, experiments: pd.DataFrame) -> float:
    return experiments.y.min()


random_results = run(
    Himmelblau(),
    strategy_factory=lambda domain: strategies.map(RandomStrategy(domain=domain)),
    n_iterations=50 if not SMOKE_TEST else 1,
    metric=best,
    initial_sampler=sample,
    n_runs=1,
    n_procs=1,
)

  0%|          | 0/1 [00:00<?, ?it/s]Run 0:   0%|          | 0/1 [00:00<?, ?it/s]Run 0:   0%|          | 0/1 [00:00<?, ?it/s, Current Best:=7.784]Run 0: 100%|██████████| 1/1 [00:00<00:00, 75.43it/s, Current Best:=7.784]

SOBO (GPEI) Optimization

def strategy_factory(domain: Domain):
    data_model = SoboStrategy(domain=domain, acquisition_function=qLogEI())
    return strategies.map(data_model)


bo_results = run(
    Himmelblau(),
    strategy_factory=strategy_factory,
    n_iterations=50 if not SMOKE_TEST else 1,
    metric=best,
    initial_sampler=sample,
    n_runs=1,
    n_procs=1,
)

  0%|          | 0/1 [00:00<?, ?it/s]Run 0:   0%|          | 0/1 [00:00<?, ?it/s]Run 0:   0%|          | 0/1 [00:00<?, ?it/s, Current Best:=20.045]Run 0: 100%|██████████| 1/1 [00:00<00:00,  2.73it/s, Current Best:=20.045]Run 0: 100%|██████████| 1/1 [00:00<00:00,  2.72it/s, Current Best:=20.045]