cotengra.experimental.hyper_pymoo¶

Hyper optimization using pymoo single-objective algorithms.

This backend currently supports serial optimization only. Pymoo ask/tell algorithms operate on generations/batches rather than individual trials, so the integration buffers one full batch at a time and feeds it back when all batch members have been evaluated.

Classes¶

`HyperPymooSampler`	Per-method ask/tell wrapper around a pymoo algorithm.
`PymooOptLib`	Hyper-optimization using pymoo algorithms with LCB method choice.

Functions¶

_get_pymoo_algorithm(name)

Module Contents¶

cotengra.experimental.hyper_pymoo._get_pymoo_algorithm(name)¶

class cotengra.experimental.hyper_pymoo.HyperPymooSampler(space, sampler='de', sampler_opts=None, exponential_param_power=None, seed=None)¶

Per-method ask/tell wrapper around a pymoo algorithm.

_np¶

_Evaluator¶

_StaticProblem¶

params¶

_ndim¶

_problem¶

algorithm¶

_trial_counter = 0¶

_active_batch = None¶

ask()¶

tell(trial_number, score)¶

class cotengra.experimental.hyper_pymoo.PymooOptLib¶

Bases: cotengra.experimental.hyper.HyperOptLib

Hyper-optimization using pymoo algorithms with LCB method choice.

setup(methods, space, optimizer=None, sampler='de', sampler_opts=None, method_exploration=1.0, method_temperature=1.0, exponential_param_power=None, seed=None, **kwargs)¶

get_setting()¶

report_result(setting, trial, score)¶