optimise

optimise(
    executor,
    ranges,
    outcome_col=None,
    n_start=20,
    n_step=10,
    max_iter=5,
    max_hours=None,
    max_gap_ratio=0.2,
    min_r2_score=0.5,
    max_avg_cv=0.15,
    max_max_cv=0.3,
)

Runs the automated Active Learning loop (Initialize -> Execute -> Diagnose -> Refine).

Parameters

Name	Type	Description	Default
executor	Executor | str	The solver adapter to use (e.g., PythonExecutor, MatlabExecutor). Accepts a legacy command string for backward compatibility.	required
ranges	Dict	Input bounds, e.g. {“Length”: (0, 10)}.	required
n_start	int	Initial sample size (only if data is empty).	20
n_step	int	Batch size for refinement.	10
max_iter	int	Max refinement loops.	5
max_hours	float	Physical time limit in hours to safely stop the loop.	None

Examples

from digiqual.core import SimulationStudy
from digiqual.executors import PythonExecutor

# 1. Define the variable ranges
ranges = {"Length": (0, 10), "Angle": (-45, 45)}
study = SimulationStudy(input_cols=["Length", "Angle"], outcome_col="Signal")

# 2. Define a simple Python solver
def my_solver(row):
    return row['Length'] * 2 + row['Angle']

my_exec = PythonExecutor(solver_func=my_solver, outcome_col="Signal")

# 3. Run the automated loop
study.optimise(
    executor=my_exec,
    ranges=ranges,
    max_iter=3
)

# 4. View the results
_ = study.pod(poi_col="Length", threshold=4.0)
study.visualise()