Bayesian exploration from YAML¶

In [1]:

from xopt import Xopt

# set values if testing
import os

import warnings

warnings.filterwarnings("ignore")


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

YAML = """
generator:
    name: bayesian_exploration

evaluator:
    function: xopt.resources.test_functions.tnk.evaluate_TNK

vocs:
    variables:
        x1: [0, 3.14159]
        x2: [0, 3.14159]
    observables: [y1]
    constraints:
        c1: [GREATER_THAN, 0]
        c2: [LESS_THAN, 0.5]
    constants: {a: dummy_constant}

"""

from xopt import Xopt # set values if testing import os import warnings warnings.filterwarnings("ignore") SMOKE_TEST = os.environ.get("SMOKE_TEST") YAML = """ generator: name: bayesian_exploration evaluator: function: xopt.resources.test_functions.tnk.evaluate_TNK vocs: variables: x1: [0, 3.14159] x2: [0, 3.14159] observables: [y1] constraints: c1: [GREATER_THAN, 0] c2: [LESS_THAN, 0.5] constants: {a: dummy_constant} """

In [2]:

X = Xopt.from_yaml(YAML)

# for testing purposes only
if SMOKE_TEST:
    X.generator.numerical_optimizer.n_restarts = 1
    X.generator.n_monte_carlo_samples = 1

X

X = Xopt.from_yaml(YAML) # for testing purposes only if SMOKE_TEST: X.generator.numerical_optimizer.n_restarts = 1 X.generator.n_monte_carlo_samples = 1 X

Out[2]:

            Xopt
________________________________
Version: 2.6.4.dev6+g92995519.d20250623
Data size: 0
Config as YAML:
dump_file: null
evaluator:
  function: xopt.resources.test_functions.tnk.evaluate_TNK
  function_kwargs:
    raise_probability: 0
    random_sleep: 0
    sleep: 0
  max_workers: 1
  vectorized: false
generator:
  computation_time: null
  custom_objective: null
  fixed_features: null
  gp_constructor:
    covar_modules: {}
    custom_noise_prior: null
    mean_modules: {}
    name: standard
    trainable_mean_keys: []
    transform_inputs: true
    use_cached_hyperparameters: false
    use_low_noise_prior: false
  max_travel_distances: null
  model: null
  n_candidates: 1
  n_interpolate_points: null
  n_monte_carlo_samples: 128
  name: bayesian_exploration
  numerical_optimizer:
    max_iter: 2000
    max_time: 5.0
    n_restarts: 20
    name: LBFGS
  supports_batch_generation: true
  supports_constraints: true
  turbo_controller: null
  use_cuda: false
max_evaluations: null
serialize_inline: false
serialize_torch: false
strict: true
vocs:
  constants:
    a: dummy_constant
  constraints:
    c1:
    - GREATER_THAN
    - 0.0
    c2:
    - LESS_THAN
    - 0.5
  objectives: {}
  observables:
  - y1
  variables:
    x1:
    - 0.0
    - 3.14159
    x2:
    - 0.0
    - 3.14159

In [3]:

X.random_evaluate(5)

for i in range(5):
    print(f"step {i}")
    X.step()

X.random_evaluate(5) for i in range(5): print(f"step {i}") X.step()

step 0

step 1

step 2

step 3

step 4

In [4]:

print(X.data)

print(X.data)

         x1        x2               a        y1        y2        c1        c2  \
0  0.988666  2.662217  dummy_constant  0.988666  2.662217  6.981981  4.913978   
1  1.257466  3.083156  dummy_constant  1.257466  3.083156  9.987454  7.246452   
2  1.727497  2.077784  dummy_constant  1.727497  2.077784  6.291239  3.996150   
3  1.110676  0.927276  dummy_constant  1.110676  0.927276  1.080005  0.555490   
4  0.053590  2.891564  dummy_constant  0.053590  2.891564  7.268378  5.918860   
5  3.141590  0.275508  dummy_constant  3.141590  0.275508  8.928454  7.028394   
6  0.000000  0.482526  dummy_constant  0.000000  0.482526 -0.867168  0.250305   
7  0.611895  1.204981  dummy_constant  0.611895  1.204981  0.793421  0.509519   
8  0.616158  1.105087  dummy_constant  0.616158  1.105087  0.628902  0.379623   
9  1.014872  0.808423  dummy_constant  1.014872  0.808423  0.706620  0.360218   

   xopt_runtime  xopt_error  
0      0.000381       False  
1      0.000136       False  
2      0.000129       False  
3      0.000126       False  
4      0.000127       False  
5      0.000153       False  
6      0.000149       False  
7      0.000151       False  
8      0.000150       False  
9      0.000151       False  

In [5]:

# plot results
ax = X.data.plot("x1", "x2")
ax.set_aspect("equal")

# plot results ax = X.data.plot("x1", "x2") ax.set_aspect("equal")

In [6]:

fig, ax = X.generator.visualize_model(show_feasibility=True, n_grid=100)

fig, ax = X.generator.visualize_model(show_feasibility=True, n_grid=100)