File indexing completed on 2026-11-09 09:29:52
0001 """Minimal Ax optimizer-only DTLZ2 example."""
0002
0003 from __future__ import annotations
0004
0005 import sys
0006 from pathlib import Path
0007 from typing import List
0008
0009 REPO_ROOT = Path(__file__).resolve().parents[3]
0010 if str(REPO_ROOT) not in sys.path:
0011 sys.path.insert(0, str(REPO_ROOT))
0012
0013 from aid2e.utilities import build_optimizer_from_config
0014 from aid2e.utilities.configurations import load_config
0015 from examples.dtlz2.dtlz2_evaluator import design_vector, dtlz2_both_objectives
0016
0017
0018 DEFAULT_CONFIG_PATH = REPO_ROOT / "examples" / "dtlz2" / "configurations" / "dtlz2_ax_joblib.yml"
0019
0020
0021 def main(argv: List[str]) -> int:
0022 config_file = Path(argv[1]).resolve() if len(argv) >= 2 else DEFAULT_CONFIG_PATH
0023 config = load_config(str(config_file))
0024 optimizer_config = config.optimizer.parse_algorithm_params()
0025 if optimizer_config is None:
0026 raise RuntimeError("No registered optimizer config model found for Ax example.")
0027
0028 optimizer = build_optimizer_from_config(config.problem, config.optimizer)
0029
0030 print("Backend: ax")
0031 print(f"Config: {config_file}")
0032 print(f"Objectives: {[objective.name for objective in config.problem.objectives]}")
0033 print(f"Design variables: {len(config.problem.design_config.get_flat_parameters())}")
0034 print(f"Generator: {optimizer_config.generator}")
0035 print(f"Initial samples: {optimizer_config.n_initial_samples}")
0036 print(f"Evaluation budget: {optimizer_config.n_iterations}")
0037 print(f"Batch size: {optimizer_config.batch_size}")
0038 print(
0039 f"\n{'Trial':<6} {'Phase':<10} {'x1':<10} {'x2':<10} {'x3':<10} "
0040 f"{'f1':<12} {'f2':<12}"
0041 )
0042 print("-" * 80)
0043
0044 remaining_init = min(optimizer_config.n_initial_samples, optimizer_config.n_iterations)
0045 init_batch = 0
0046 while remaining_init > 0:
0047 init_batch += 1
0048 current_batch = min(optimizer_config.batch_size, remaining_init)
0049 candidates = optimizer.suggest_candidates(n_candidates=current_batch)
0050 start_index = len(optimizer.get_trials()) - len(candidates)
0051 for offset, parameters in enumerate(candidates):
0052 trial_index = start_index + offset
0053 metrics = dtlz2_both_objectives(design_vector(parameters))
0054 optimizer.update_with_results(
0055 trial_index=trial_index,
0056 parameters=parameters,
0057 metrics=metrics,
0058 )
0059 vector = design_vector(parameters)
0060 print(
0061 f"{trial_index + 1:<6} {'init':<10} "
0062 f"{vector[0]:<10.4f} {vector[1]:<10.4f} {vector[2]:<10.4f} "
0063 f"{metrics['f1']:<12.6f} {metrics['f2']:<12.6f}"
0064 )
0065 remaining_init -= current_batch
0066
0067 for iteration, start in enumerate(range(
0068 len(optimizer.get_trials()), optimizer_config.n_iterations, optimizer_config.batch_size
0069 )):
0070 candidates = optimizer.suggest_candidates(
0071 n_candidates=min(optimizer_config.batch_size, optimizer_config.n_iterations - start)
0072 )
0073 start_index = len(optimizer.get_trials()) - len(candidates)
0074 for offset, parameters in enumerate(candidates):
0075 trial_index = start_index + offset
0076 metrics = dtlz2_both_objectives(design_vector(parameters))
0077 optimizer.update_with_results(
0078 trial_index=trial_index,
0079 parameters=parameters,
0080 metrics=metrics,
0081 )
0082 vector = design_vector(parameters)
0083 print(
0084 f"{trial_index + 1:<6} {f'iter-{iteration + 1}':<10} "
0085 f"{vector[0]:<10.4f} {vector[1]:<10.4f} {vector[2]:<10.4f} "
0086 f"{metrics['f1']:<12.6f} {metrics['f2']:<12.6f}"
0087 )
0088
0089 results = optimizer.get_optimization_results()
0090 pareto_front = optimizer.get_pareto_front()
0091
0092 print("\nSummary")
0093 print(f"Trials recorded: {results['n_trials']}")
0094 print(f"Pareto points: {len(pareto_front)}")
0095 return 0
0096
0097
0098 if __name__ == "__main__":
0099 raise SystemExit(main(sys.argv))
