remove with_count parameter (act like it's True)
This commit is contained in:
parent
ac39efe00a
commit
6778965d70
10 changed files with 93 additions and 106 deletions
|
|
@ -3,9 +3,9 @@ from .random import another_random_cube
|
|||
from ..utilities import wrap_untrustworthy, final
|
||||
|
||||
|
||||
def dlib_cube(objective, n_trials, n_dim, with_count):
|
||||
def dlib_cube(objective, n_trials, n_dim):
|
||||
if n_dim > 35:
|
||||
return another_random_cube(objective, n_trials, n_dim, with_count)
|
||||
return another_random_cube(objective, n_trials, n_dim)
|
||||
|
||||
_objective = wrap_untrustworthy(objective, n_trials)
|
||||
|
||||
|
|
@ -13,5 +13,4 @@ def dlib_cube(objective, n_trials, n_dim, with_count):
|
|||
return _objective(list(args))
|
||||
|
||||
find_min_global(__objective, [0.0] * n_dim, [1.0] * n_dim, n_trials)
|
||||
fopt, xopt, feval_count = _objective(final)
|
||||
return (fopt, xopt, feval_count) if with_count else (fopt, xopt)
|
||||
return _objective(final)
|
||||
|
|
|
|||
|
|
@ -43,7 +43,7 @@ def make_evolopy(optimizer_name, popsize=None):
|
|||
}
|
||||
optimizer = optimizers[optimizer_name.upper()]
|
||||
|
||||
def f(objective, n_trials, n_dim, with_count):
|
||||
def f(objective, n_trials, n_dim):
|
||||
# ps = n_dim if popsize is None else popsize
|
||||
# ps = max(4, ps) if optimizer_name.upper() == "DE" else ps
|
||||
ps = int(4 + 3 * np.log(n_dim)) if popsize is None else popsize
|
||||
|
|
@ -58,8 +58,7 @@ def make_evolopy(optimizer_name, popsize=None):
|
|||
except ExhaustedTrialsError:
|
||||
pass
|
||||
|
||||
fopt, xopt, feval_count = _objective(final)
|
||||
return (fopt, xopt, feval_count) if with_count else (fopt, xopt)
|
||||
return _objective(final)
|
||||
|
||||
name = f"evolopy_{optimizer.__name__.lower()}"
|
||||
name += f"_auto" if popsize is None else f"_ps{popsize:02}"
|
||||
|
|
|
|||
|
|
@ -21,13 +21,12 @@ def _fix_logging(original_function):
|
|||
def make_biteopt(depth=1):
|
||||
from fcmaes.optimizer import Bite_cpp
|
||||
|
||||
def f(objective, n_trials, n_dim, with_count):
|
||||
def f(objective, n_trials, n_dim):
|
||||
_objective = wrap_untrustworthy(objective, n_trials)
|
||||
bounds = Bounds([0.0] * n_dim, [1.0] * n_dim)
|
||||
optim = Bite_cpp(max_evaluations=n_trials, M=depth)
|
||||
_xopt, _fopt, _feval_count = optim.minimize(_objective, bounds)
|
||||
fopt, xopt, feval_count = _objective(final)
|
||||
return (fopt, xopt, feval_count) if with_count else (fopt, xopt)
|
||||
return _objective(final)
|
||||
|
||||
name = f"fcmaes_biteopt_d{depth}"
|
||||
f.__name__ = name + "_cube"
|
||||
|
|
@ -38,15 +37,14 @@ def make_biteopt(depth=1):
|
|||
def make_csma(isigma=3):
|
||||
from fcmaes.optimizer import Csma_cpp
|
||||
|
||||
def f(objective, n_trials, n_dim, with_count):
|
||||
def f(objective, n_trials, n_dim):
|
||||
_objective = wrap_untrustworthy(objective, n_trials)
|
||||
bounds = Bounds([0.0] * n_dim, [1.0] * n_dim)
|
||||
optim = Csma_cpp(max_evaluations=n_trials)
|
||||
_xopt, _fopt, _feval_count = optim.minimize(
|
||||
_objective, bounds, sdevs=1 / isigma
|
||||
)
|
||||
fopt, xopt, feval_count = _objective(final)
|
||||
return (fopt, xopt, feval_count) if with_count else (fopt, xopt)
|
||||
return _objective(final)
|
||||
|
||||
name = f"fcmaes_csma_is{isigma:02}"
|
||||
f.__name__ = name + "_cube"
|
||||
|
|
@ -57,7 +55,7 @@ def make_csma(isigma=3):
|
|||
def make_fcmaes(popsize=None):
|
||||
from fcmaes.optimizer import Cma_cpp
|
||||
|
||||
def f(objective, n_trials, n_dim, with_count):
|
||||
def f(objective, n_trials, n_dim):
|
||||
_objective = wrap_untrustworthy(objective, n_trials)
|
||||
bounds = Bounds([0.0] * n_dim, [1.0] * n_dim)
|
||||
ps = int(4 + 3 * np.log(n_dim)) if popsize is None else popsize
|
||||
|
|
@ -69,8 +67,7 @@ def make_fcmaes(popsize=None):
|
|||
delayed_update=False,
|
||||
)
|
||||
_xopt, _fopt, _feval_count = optim.minimize(_objective, bounds, sdevs=1 / 3)
|
||||
fopt, xopt, feval_count = _objective(final)
|
||||
return (fopt, xopt, feval_count) if with_count else (fopt, xopt)
|
||||
return _objective(final)
|
||||
|
||||
name = "fcmaes_cma"
|
||||
name += f"_auto" if popsize is None else f"_ps{popsize}"
|
||||
|
|
@ -82,14 +79,13 @@ def make_fcmaes(popsize=None):
|
|||
def make_crfmnes(popsize=None):
|
||||
from fcmaes.optimizer import Crfmnes_cpp
|
||||
|
||||
def f(objective, n_trials, n_dim, with_count):
|
||||
def f(objective, n_trials, n_dim):
|
||||
_objective = wrap_untrustworthy(objective, n_trials)
|
||||
bounds = Bounds([0.0] * n_dim, [1.0] * n_dim)
|
||||
ps = int(4 + 3 * np.log(n_dim)) if popsize is None else popsize
|
||||
optim = Crfmnes_cpp(max_evaluations=n_trials, popsize=ps)
|
||||
_xopt, _fopt, _feval_count = optim.minimize(_objective, bounds, sdevs=1 / 3)
|
||||
fopt, xopt, feval_count = _objective(final)
|
||||
return (fopt, xopt, feval_count) if with_count else (fopt, xopt)
|
||||
return _objective(final)
|
||||
|
||||
name = "fcmaes_crfmnes"
|
||||
name += f"_auto" if popsize is None else f"_ps{popsize}"
|
||||
|
|
@ -102,13 +98,12 @@ _broken = """
|
|||
def make_lde(popsize=None):
|
||||
from fcmaes.optimizer import LDe_cpp
|
||||
|
||||
def f(objective, n_trials, n_dim, with_count):
|
||||
def f(objective, n_trials, n_dim):
|
||||
_objective = wrap_untrustworthy(objective, n_trials)
|
||||
bounds = Bounds([0.0] * n_dim, [1.0] * n_dim)
|
||||
optim = LDe_cpp(max_evaluations=n_trials, popsize=popsize)
|
||||
_xopt, _fopt, _feval_count = optim.minimize(_objective, bounds)
|
||||
fopt, xopt, feval_count = _objective(final)
|
||||
return (fopt, xopt, feval_count) if with_count else (fopt, xopt)
|
||||
return _objective(final)
|
||||
|
||||
name = "fcmaes_lde"
|
||||
name += f"_auto" if popsize is None else f"_ps{popsize}"
|
||||
|
|
@ -121,13 +116,12 @@ def make_lde(popsize=None):
|
|||
def make_da(local=True):
|
||||
from fcmaes.optimizer import Da_cpp
|
||||
|
||||
def f(objective, n_trials, n_dim, with_count):
|
||||
def f(objective, n_trials, n_dim):
|
||||
_objective = wrap_untrustworthy(objective, n_trials)
|
||||
bounds = Bounds([0.0] * n_dim, [1.0] * n_dim)
|
||||
optim = Da_cpp(max_evaluations=n_trials, use_local_search=local)
|
||||
_xopt, _fopt, _feval_count = optim.minimize(_objective, bounds)
|
||||
fopt, xopt, feval_count = _objective(final)
|
||||
return (fopt, xopt, feval_count) if with_count else (fopt, xopt)
|
||||
return _objective(final)
|
||||
|
||||
name = "fcmaes_da"
|
||||
if local:
|
||||
|
|
@ -140,13 +134,12 @@ def make_da(local=True):
|
|||
def make_gclde(popsize=None):
|
||||
from fcmaes.optimizer import GCLDE_cpp
|
||||
|
||||
def f(objective, n_trials, n_dim, with_count):
|
||||
def f(objective, n_trials, n_dim):
|
||||
_objective = wrap_untrustworthy(objective, n_trials)
|
||||
bounds = Bounds([0.0] * n_dim, [1.0] * n_dim)
|
||||
optim = GCLDE_cpp(max_evaluations=n_trials, popsize=popsize)
|
||||
_xopt, _fopt, _feval_count = optim.minimize(_objective, bounds)
|
||||
fopt, xopt, feval_count = _objective(final)
|
||||
return (fopt, xopt, feval_count) if with_count else (fopt, xopt)
|
||||
return _objective(final)
|
||||
|
||||
name = "fcmaes_gclde"
|
||||
name += f"_auto" if popsize is None else f"_ps{popsize}"
|
||||
|
|
@ -158,13 +151,12 @@ def make_gclde(popsize=None):
|
|||
def make_lclde(popsize=None):
|
||||
from fcmaes.optimizer import LCLDE_cpp
|
||||
|
||||
def f(objective, n_trials, n_dim, with_count):
|
||||
def f(objective, n_trials, n_dim):
|
||||
_objective = wrap_untrustworthy(objective, n_trials)
|
||||
bounds = Bounds([0.0] * n_dim, [1.0] * n_dim)
|
||||
optim = LCLDE_cpp(max_evaluations=n_trials, popsize=popsize)
|
||||
_xopt, _fopt, _feval_count = optim.minimize(_objective, bounds)
|
||||
fopt, xopt, feval_count = _objective(final)
|
||||
return (fopt, xopt, feval_count) if with_count else (fopt, xopt)
|
||||
return _objective(final)
|
||||
|
||||
name = "fcmaes_lclde"
|
||||
name += f"_auto" if popsize is None else f"_ps{popsize}"
|
||||
|
|
|
|||
|
|
@ -303,7 +303,7 @@ assert ol.RSLSQP is ol.RSQP, "weirdness is gone, please adjust accordingly"
|
|||
assert ol.SLSQP is ol.SQP, "weirdness is gone, please adjust accordingly"
|
||||
|
||||
|
||||
def nevergrad_cube_factory(optimizer, objective, n_trials, n_dim, with_count):
|
||||
def nevergrad_cube_factory(optimizer, objective, n_trials, n_dim):
|
||||
instrument = ng.p.Array(lower=0, upper=1, shape=(n_dim,)) # better sigma
|
||||
# ev.RBO, ev.QRBO, ev.MidQRBO, and ev.LBO still complain, though:
|
||||
# /home/py/.local/lib/python3.10/site-packages/nevergrad/parametrization/_datalayers.py:107: NevergradRuntimeWarning: Bounds are 1.0 sigma away from each other at the closest, you should aim for at least 3 for better quality.
|
||||
|
|
@ -326,7 +326,7 @@ def nevergrad_cube_factory(optimizer, objective, n_trials, n_dim, with_count):
|
|||
best_x = recommendation.value
|
||||
best_val = cube_objective(best_x) # don't trust recommendation.loss
|
||||
|
||||
return (best_val, best_x, feval_count) if with_count else (best_val, best_x)
|
||||
return best_val, best_x, feval_count
|
||||
|
||||
|
||||
def named_optimizer(optimizer, experimental=False):
|
||||
|
|
|
|||
|
|
@ -47,7 +47,7 @@ NLOPTIMIZERS = {
|
|||
}
|
||||
|
||||
|
||||
def nlopt_cube_factory(objective, n_trials, n_dim, with_count, method):
|
||||
def nlopt_cube_factory(objective, n_trials, n_dim, method):
|
||||
optim = NLOPTIMIZERS[method]
|
||||
|
||||
feval_count = 0
|
||||
|
|
@ -83,110 +83,109 @@ def nlopt_cube_factory(objective, n_trials, n_dim, with_count, method):
|
|||
|
||||
assert best_so_far is not None, optimizer.__name__
|
||||
fopt, xopt = best_so_far
|
||||
|
||||
return (fopt, xopt, feval_count) if with_count else (fopt, xopt)
|
||||
return fopt, xopt, feval_count
|
||||
|
||||
|
||||
def nlopt_ags_cube(objective, n_trials, n_dim, with_count):
|
||||
return nlopt_cube_factory(objective, n_trials, n_dim, with_count, "gn_ags")
|
||||
def nlopt_ags_cube(objective, n_trials, n_dim):
|
||||
return nlopt_cube_factory(objective, n_trials, n_dim, "gn_ags")
|
||||
|
||||
|
||||
def nlopt_crs2_lm_cube(objective, n_trials, n_dim, with_count):
|
||||
return nlopt_cube_factory(objective, n_trials, n_dim, with_count, "gn_crs2_lm")
|
||||
def nlopt_crs2_lm_cube(objective, n_trials, n_dim):
|
||||
return nlopt_cube_factory(objective, n_trials, n_dim, "gn_crs2_lm")
|
||||
|
||||
|
||||
def nlopt_direct_cube(objective, n_trials, n_dim, with_count):
|
||||
return nlopt_cube_factory(objective, n_trials, n_dim, with_count, "gn_direct")
|
||||
def nlopt_direct_cube(objective, n_trials, n_dim):
|
||||
return nlopt_cube_factory(objective, n_trials, n_dim, "gn_direct")
|
||||
|
||||
|
||||
def nlopt_direct_l_cube(objective, n_trials, n_dim, with_count):
|
||||
return nlopt_cube_factory(objective, n_trials, n_dim, with_count, "gn_direct_l")
|
||||
def nlopt_direct_l_cube(objective, n_trials, n_dim):
|
||||
return nlopt_cube_factory(objective, n_trials, n_dim, "gn_direct_l")
|
||||
|
||||
|
||||
def nlopt_direct_l_noscal_cube(objective, n_trials, n_dim, with_count):
|
||||
def nlopt_direct_l_noscal_cube(objective, n_trials, n_dim):
|
||||
return nlopt_cube_factory(
|
||||
objective, n_trials, n_dim, with_count, "gn_direct_l_noscal"
|
||||
objective, n_trials, n_dim, "gn_direct_l_noscal"
|
||||
)
|
||||
|
||||
|
||||
def nlopt_direct_lr_cube(objective, n_trials, n_dim, with_count):
|
||||
def nlopt_direct_lr_cube(objective, n_trials, n_dim):
|
||||
return nlopt_cube_factory(
|
||||
objective, n_trials, n_dim, with_count, "gn_direct_l_rand"
|
||||
objective, n_trials, n_dim, "gn_direct_l_rand"
|
||||
)
|
||||
|
||||
|
||||
def nlopt_direct_lr_noscal_cube(objective, n_trials, n_dim, with_count):
|
||||
def nlopt_direct_lr_noscal_cube(objective, n_trials, n_dim):
|
||||
return nlopt_cube_factory(
|
||||
objective, n_trials, n_dim, with_count, "gn_direct_l_rand_noscal"
|
||||
objective, n_trials, n_dim, "gn_direct_l_rand_noscal"
|
||||
)
|
||||
|
||||
|
||||
def nlopt_direct_noscal_cube(objective, n_trials, n_dim, with_count):
|
||||
def nlopt_direct_noscal_cube(objective, n_trials, n_dim):
|
||||
return nlopt_cube_factory(
|
||||
objective, n_trials, n_dim, with_count, "gn_direct_noscal"
|
||||
objective, n_trials, n_dim, "gn_direct_noscal"
|
||||
)
|
||||
|
||||
|
||||
def nlopt_esch_cube(objective, n_trials, n_dim, with_count):
|
||||
return nlopt_cube_factory(objective, n_trials, n_dim, with_count, "gn_esch")
|
||||
def nlopt_esch_cube(objective, n_trials, n_dim):
|
||||
return nlopt_cube_factory(objective, n_trials, n_dim, "gn_esch")
|
||||
|
||||
|
||||
def nlopt_isres_cube(objective, n_trials, n_dim, with_count):
|
||||
return nlopt_cube_factory(objective, n_trials, n_dim, with_count, "gn_isres")
|
||||
def nlopt_isres_cube(objective, n_trials, n_dim):
|
||||
return nlopt_cube_factory(objective, n_trials, n_dim, "gn_isres")
|
||||
|
||||
|
||||
def nlopt_mlsl_cube(objective, n_trials, n_dim, with_count):
|
||||
return nlopt_cube_factory(objective, n_trials, n_dim, with_count, "gn_mlsl")
|
||||
def nlopt_mlsl_cube(objective, n_trials, n_dim):
|
||||
return nlopt_cube_factory(objective, n_trials, n_dim, "gn_mlsl")
|
||||
|
||||
|
||||
def nlopt_mlsl_lds_cube(objective, n_trials, n_dim, with_count):
|
||||
return nlopt_cube_factory(objective, n_trials, n_dim, with_count, "gn_mlsl_lds")
|
||||
def nlopt_mlsl_lds_cube(objective, n_trials, n_dim):
|
||||
return nlopt_cube_factory(objective, n_trials, n_dim, "gn_mlsl_lds")
|
||||
|
||||
|
||||
def nlopt_orig_direct_cube(objective, n_trials, n_dim, with_count):
|
||||
return nlopt_cube_factory(objective, n_trials, n_dim, with_count, "gn_orig_direct")
|
||||
def nlopt_orig_direct_cube(objective, n_trials, n_dim):
|
||||
return nlopt_cube_factory(objective, n_trials, n_dim, "gn_orig_direct")
|
||||
|
||||
|
||||
def nlopt_orig_direct_l_cube(objective, n_trials, n_dim, with_count):
|
||||
def nlopt_orig_direct_l_cube(objective, n_trials, n_dim):
|
||||
return nlopt_cube_factory(
|
||||
objective, n_trials, n_dim, with_count, "gn_orig_direct_l"
|
||||
objective, n_trials, n_dim, "gn_orig_direct_l"
|
||||
)
|
||||
|
||||
|
||||
def nlopt_auglag_cube(objective, n_trials, n_dim, with_count):
|
||||
return nlopt_cube_factory(objective, n_trials, n_dim, with_count, "ln_auglag")
|
||||
def nlopt_auglag_cube(objective, n_trials, n_dim):
|
||||
return nlopt_cube_factory(objective, n_trials, n_dim, "ln_auglag")
|
||||
|
||||
|
||||
def nlopt_auglag_eq_cube(objective, n_trials, n_dim, with_count):
|
||||
return nlopt_cube_factory(objective, n_trials, n_dim, with_count, "ln_auglag_eq")
|
||||
def nlopt_auglag_eq_cube(objective, n_trials, n_dim):
|
||||
return nlopt_cube_factory(objective, n_trials, n_dim, "ln_auglag_eq")
|
||||
|
||||
|
||||
def nlopt_bobyqa_cube(objective, n_trials, n_dim, with_count):
|
||||
return nlopt_cube_factory(objective, n_trials, n_dim, with_count, "ln_bobyqa")
|
||||
def nlopt_bobyqa_cube(objective, n_trials, n_dim):
|
||||
return nlopt_cube_factory(objective, n_trials, n_dim, "ln_bobyqa")
|
||||
|
||||
|
||||
def nlopt_cobyla_cube(objective, n_trials, n_dim, with_count):
|
||||
return nlopt_cube_factory(objective, n_trials, n_dim, with_count, "ln_cobyla")
|
||||
def nlopt_cobyla_cube(objective, n_trials, n_dim):
|
||||
return nlopt_cube_factory(objective, n_trials, n_dim, "ln_cobyla")
|
||||
|
||||
|
||||
def nlopt_neldermead_cube(objective, n_trials, n_dim, with_count):
|
||||
return nlopt_cube_factory(objective, n_trials, n_dim, with_count, "ln_neldermead")
|
||||
def nlopt_neldermead_cube(objective, n_trials, n_dim):
|
||||
return nlopt_cube_factory(objective, n_trials, n_dim, "ln_neldermead")
|
||||
|
||||
|
||||
def nlopt_newuoa_cube(objective, n_trials, n_dim, with_count):
|
||||
return nlopt_cube_factory(objective, n_trials, n_dim, with_count, "ln_newuoa")
|
||||
def nlopt_newuoa_cube(objective, n_trials, n_dim):
|
||||
return nlopt_cube_factory(objective, n_trials, n_dim, "ln_newuoa")
|
||||
|
||||
|
||||
def nlopt_newuoa_bound_cube(objective, n_trials, n_dim, with_count):
|
||||
return nlopt_cube_factory(objective, n_trials, n_dim, with_count, "ln_newuoa_bound")
|
||||
def nlopt_newuoa_bound_cube(objective, n_trials, n_dim):
|
||||
return nlopt_cube_factory(objective, n_trials, n_dim, "ln_newuoa_bound")
|
||||
|
||||
|
||||
def nlopt_praxis_cube(objective, n_trials, n_dim, with_count):
|
||||
return nlopt_cube_factory(objective, n_trials, n_dim, with_count, "ln_praxis")
|
||||
def nlopt_praxis_cube(objective, n_trials, n_dim):
|
||||
return nlopt_cube_factory(objective, n_trials, n_dim, "ln_praxis")
|
||||
|
||||
|
||||
def nlopt_sbplx_cube(objective, n_trials, n_dim, with_count):
|
||||
return nlopt_cube_factory(objective, n_trials, n_dim, with_count, "ln_sbplx")
|
||||
def nlopt_sbplx_cube(objective, n_trials, n_dim):
|
||||
return nlopt_cube_factory(objective, n_trials, n_dim, "ln_sbplx")
|
||||
|
||||
|
||||
NLOPT_OPTIMIZERS = [
|
||||
|
|
|
|||
|
|
@ -6,7 +6,7 @@ import numpy as np
|
|||
def make_birect(deepness=23, *, longest=False, pruning=False):
|
||||
from ..internal.birect import birect
|
||||
|
||||
def f(objective, n_trials, n_dim, with_count):
|
||||
def f(objective, n_trials, n_dim):
|
||||
feval_count = 0
|
||||
|
||||
def _objective(x):
|
||||
|
|
@ -24,7 +24,7 @@ def make_birect(deepness=23, *, longest=False, pruning=False):
|
|||
pruning=pruning,
|
||||
F=np.float64,
|
||||
)
|
||||
return (fopt, xopt, feval_count) if with_count else (fopt, xopt)
|
||||
return fopt, xopt, feval_count
|
||||
|
||||
name = f"birect{deepness:02}"
|
||||
name += "_longest" if longest else ""
|
||||
|
|
@ -39,7 +39,7 @@ def make_soo(deepness=None, *, K=3):
|
|||
assert K >= 2
|
||||
from ..internal.soo import soo
|
||||
|
||||
def f(objective, n_trials, n_dim, with_count):
|
||||
def f(objective, n_trials, n_dim):
|
||||
feval_count = 0
|
||||
|
||||
def _objective(x):
|
||||
|
|
@ -51,7 +51,7 @@ def make_soo(deepness=None, *, K=3):
|
|||
_objective, np.full(n_dim, 0.5), 0.5, n_trials, K=K, h_max=deepness
|
||||
)
|
||||
fopt = history[-1]
|
||||
return (fopt, xopt, feval_count) if with_count else (fopt, xopt)
|
||||
return fopt, xopt, feval_count
|
||||
|
||||
name = f"soo{deepness}_k{K}"
|
||||
f.__name__ = name + "_cube"
|
||||
|
|
@ -63,7 +63,7 @@ def make_mercury(
|
|||
):
|
||||
from ..internal.hg import minimize as hg
|
||||
|
||||
def f(objective, n_trials, n_dim, with_count):
|
||||
def f(objective, n_trials, n_dim):
|
||||
_objective = wrap_untrustworthy(objective, n_trials, bounding=bounding)
|
||||
|
||||
init = (0.5,) * n_dim
|
||||
|
|
@ -89,8 +89,7 @@ def make_mercury(
|
|||
x = prng.uniform(size=n_dim)
|
||||
fx = _objective(x)
|
||||
|
||||
fopt, xopt, feval_count = _objective(final)
|
||||
return (fopt, xopt, feval_count) if with_count else (fopt, xopt)
|
||||
return _objective(final)
|
||||
|
||||
name = f"hg{flags:02}"
|
||||
name += f"_{bounding}" if bounding != "clip" else ""
|
||||
|
|
|
|||
|
|
@ -2,7 +2,7 @@ from ..utilities import phi
|
|||
import numpy as np
|
||||
|
||||
|
||||
def another_random_cube(objective, n_trials, n_dim, with_count, seed=None):
|
||||
def another_random_cube(objective, n_trials, n_dim, seed=None):
|
||||
prng = np.random.default_rng(seed)
|
||||
fopt = None
|
||||
xopt = None
|
||||
|
|
@ -12,10 +12,10 @@ def another_random_cube(objective, n_trials, n_dim, with_count, seed=None):
|
|||
if fopt is None or xopt is None or fx < fopt:
|
||||
fopt = fx
|
||||
xopt = x
|
||||
return (fopt, xopt, n_trials) if with_count else (fopt, xopt)
|
||||
return fopt, xopt, n_trials
|
||||
|
||||
|
||||
def quasirandom_cube(objective, n_trials, n_dim, with_count):
|
||||
def quasirandom_cube(objective, n_trials, n_dim):
|
||||
# http://extremelearning.com.au/unreasonable-effectiveness-of-quasirandom-sequences/
|
||||
magic = phi(n_dim)
|
||||
alpha = np.zeros(n_dim)
|
||||
|
|
@ -32,4 +32,4 @@ def quasirandom_cube(objective, n_trials, n_dim, with_count):
|
|||
if best_so_far is None or fx < best_so_far[0]:
|
||||
best_so_far = (fx, x)
|
||||
fopt, xopt = best_so_far
|
||||
return (fopt, xopt, n_trials) if with_count else (fopt, xopt)
|
||||
return fopt, xopt, n_trials
|
||||
|
|
|
|||
|
|
@ -23,7 +23,7 @@ def make_scipy(method, *, jacobian=None, hessian=None):
|
|||
# "trust-krylov",
|
||||
), method
|
||||
|
||||
def f(objective, n_trials, n_dim, with_count):
|
||||
def f(objective, n_trials, n_dim):
|
||||
prng = np.random.default_rng()
|
||||
_objective = wrap_untrustworthy(
|
||||
objective, n_trials, raising=True, bounding="sine"
|
||||
|
|
@ -107,8 +107,7 @@ def make_scipy(method, *, jacobian=None, hessian=None):
|
|||
# TODO: run without this, try to minimize number of attempts (i.e. list length)
|
||||
# if len(checks) >= 5: print(method, [b - a for a, b in zip(checks, checks[1:])])
|
||||
|
||||
fopt, xopt, feval_count = _objective(final)
|
||||
return (fopt, xopt, feval_count) if with_count else (fopt, xopt)
|
||||
return _objective(final)
|
||||
|
||||
name = f"scipy_{method.replace('-', '').lower()}"
|
||||
if jacobian == "2-point":
|
||||
|
|
@ -127,7 +126,7 @@ def make_scipy(method, *, jacobian=None, hessian=None):
|
|||
return f
|
||||
|
||||
|
||||
def scipy_basinhopping_cube(objective, n_trials, n_dim, with_count):
|
||||
def scipy_basinhopping_cube(objective, n_trials, n_dim):
|
||||
progress = 1e-2 # TODO: make configurable?
|
||||
# NOTE: could also callbacks to extract solutions instead of wrapping objective functions?
|
||||
|
||||
|
|
@ -151,10 +150,10 @@ def scipy_basinhopping_cube(objective, n_trials, n_dim, with_count):
|
|||
)
|
||||
fopt, xopt, feval_count = res.fun, res.x, res.nfev
|
||||
# print("success:", res.success)
|
||||
return (fopt, xopt, feval_count) if with_count else (fopt, xopt)
|
||||
return fopt, xopt, feval_count
|
||||
|
||||
|
||||
def scipy_direct_cube(objective, n_trials, n_dim, with_count):
|
||||
def scipy_direct_cube(objective, n_trials, n_dim):
|
||||
bounds = scopt.Bounds([0.0] * n_dim, [1.0] * n_dim)
|
||||
# TODO: try different values of eps. default 0.0001
|
||||
res = scopt.direct(
|
||||
|
|
@ -166,10 +165,10 @@ def scipy_direct_cube(objective, n_trials, n_dim, with_count):
|
|||
)
|
||||
fopt, xopt, feval_count = res.fun, res.x, res.nfev
|
||||
# print("success:", res.success)
|
||||
return (fopt, xopt, feval_count) if with_count else (fopt, xopt)
|
||||
return fopt, xopt, feval_count
|
||||
|
||||
|
||||
def scipy_direct_l_cube(objective, n_trials, n_dim, with_count):
|
||||
def scipy_direct_l_cube(objective, n_trials, n_dim):
|
||||
bounds = scopt.Bounds([0.0] * n_dim, [1.0] * n_dim)
|
||||
# TODO: try different values of eps. default 0.0001
|
||||
res = scopt.direct(
|
||||
|
|
@ -183,7 +182,7 @@ def scipy_direct_l_cube(objective, n_trials, n_dim, with_count):
|
|||
)
|
||||
fopt, xopt, feval_count = res.fun, res.x, res.nfev
|
||||
# print("success:", res.success)
|
||||
return (fopt, xopt, feval_count) if with_count else (fopt, xopt)
|
||||
return fopt, xopt, feval_count
|
||||
|
||||
|
||||
def make_shgo(method="cobyla", init="large", it=1, mei=False, li=False, ftol=12):
|
||||
|
|
@ -199,7 +198,7 @@ def make_shgo(method="cobyla", init="large", it=1, mei=False, li=False, ftol=12)
|
|||
# https://docs.scipy.org/doc/scipy/reference/optimize.minimize-cobyla.html
|
||||
pass
|
||||
|
||||
def f(objective, n_trials, n_dim, with_count):
|
||||
def f(objective, n_trials, n_dim):
|
||||
_objective = wrap_untrustworthy(
|
||||
objective, n_trials, bounding="clip", raising=True
|
||||
)
|
||||
|
|
@ -220,8 +219,7 @@ def make_shgo(method="cobyla", init="large", it=1, mei=False, li=False, ftol=12)
|
|||
)
|
||||
except ExhaustedTrialsError:
|
||||
pass
|
||||
fopt, xopt, feval_count = _objective(final)
|
||||
return (fopt, xopt, feval_count) if with_count else (fopt, xopt)
|
||||
return _objective(final)
|
||||
|
||||
name = f"shgo_{method}"
|
||||
if it != 1:
|
||||
|
|
|
|||
|
|
@ -380,7 +380,7 @@ def main(argv, display=True):
|
|||
note(
|
||||
f"Using {opt_name} to optimize {obj_realname} ({obj_name}) [{run}] ..."
|
||||
)
|
||||
_ = optimizer(wrapped, n_trials=n_trials, n_dim=n_dim, with_count=False)
|
||||
_ = optimizer(wrapped, n_trials=n_trials, n_dim=n_dim)
|
||||
fopt, xopt = wrapped.finish()
|
||||
result = (fopt, opt_name, wrapped.history)
|
||||
results.setdefault(obj_name, []).append(result)
|
||||
|
|
|
|||
|
|
@ -32,6 +32,7 @@ EVOLOPY_OPTIMIZERS = [
|
|||
FCMAES_OPTIMIZERS = [
|
||||
make_biteopt(1),
|
||||
make_biteopt(2),
|
||||
make_biteopt(3),
|
||||
make_csma(3),
|
||||
make_da(False), # TODO: fix local=True case.
|
||||
] + [
|
||||
|
|
|
|||
Loading…
Reference in a new issue