#! /usr/bin/env python
# Copyright (c) 2014 KU Leuven, ESAT-STADIUS
# All rights reserved.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions
# are met:
#
# 1. Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
#
# 2. Redistributions in binary form must reproduce the above copyright
# notice, this list of conditions and the following disclaimer in the
# documentation and/or other materials provided with the distribution.
#
# 3. Neither name of copyright holders nor the names of its contributors
# may be used to endorse or promote products derived from this software
# without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
# ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR
# CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
# EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
# PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
# PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
# LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
# NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
# SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
import math
import functools
from .solver_registry import register_solver
from .util import Solver, _copydoc
from . import util
_numpy_available = True
try:
import numpy as np
except ImportError:
_numpy_available = False
_deap_available = True
try:
import deap
import deap.creator
import deap.base
import deap.tools
import deap.cma
import deap.algorithms
except ImportError:
_deap_available = False
except TypeError:
# this can happen because DEAP is in Python 2
# install needs to take proper care of converting
# 2 to 3 when necessary
_deap_available = False
[docs]class CMA_ES(Solver):
"""
.. include:: /global.rst
Please refer to |cmaes| for details about this algorithm.
This solver uses an implementation available in the DEAP library [DEAP2012]_.
.. warning:: This solver has dependencies on DEAP_ and NumPy_
and will be unavailable if these are not met.
.. _DEAP: https://code.google.com/p/deap/
.. _NumPy: http://www.numpy.org
"""
def __init__(self, num_generations, sigma=1.0, Lambda=None, **kwargs):
"""blah
.. warning:: |warning-unconstrained|
"""
if not _deap_available:
raise ImportError('This solver requires DEAP but it is missing.')
if not _numpy_available:
raise ImportError('This solver requires NumPy but it is missing.')
self._num_generations = num_generations
self._start = kwargs
self._sigma = sigma
self._lambda = Lambda
@staticmethod
[docs] def suggest_from_seed(num_evals, **kwargs):
"""Verify that we can effectively make a solver.
The doctest has to be skipped from automated builds, because DEAP may not be available
and yet we want documentation to be generated.
>>> s = CMA_ES.suggest_from_seed(30, x=1.0, y=-1.0, z=2.0)
>>> solver = CMA_ES(**s) #doctest:+SKIP
"""
fertility = 4 + 3 * math.log(len(kwargs))
d = dict(kwargs)
d['num_generations'] = int(math.ceil(float(num_evals) / fertility))
# num_gen is overestimated
# this will require slightly more function evaluations than permitted by num_evals
return d
@property
def num_generations(self):
return self._num_generations
@property
def start(self):
"""Returns the starting point for CMA-ES."""
return self._start
@property
def lambda_(self):
return self._lambda
@property
def sigma(self):
return self._sigma
@_copydoc(Solver.optimize)
[docs] def optimize(self, f, maximize=True, pmap=map):
toolbox = deap.base.Toolbox()
if maximize:
fit = 1.0
else:
fit = -1.0
deap.creator.create("FitnessMax", deap.base.Fitness,
weights=(fit,))
Fit = deap.creator.FitnessMax
deap.creator.create("Individual", list,
fitness=Fit)
Individual = deap.creator.Individual
if self.lambda_:
strategy = deap.cma.Strategy(centroid=list(self.start.values()),
sigma=self.sigma, lambda_=self.lambda_)
else:
strategy = deap.cma.Strategy(centroid=list(self.start.values()),
sigma=self.sigma)
toolbox.register("generate", strategy.generate, Individual)
toolbox.register("update", strategy.update)
@functools.wraps(f)
def evaluate(individual):
return (util.score(f(**dict([(k, v)
for k, v in zip(self.start.keys(),
individual)]))),)
toolbox.register("evaluate", evaluate)
toolbox.register("map", pmap)
hof = deap.tools.HallOfFame(1)
deap.algorithms.eaGenerateUpdate(toolbox=toolbox,
ngen=self._num_generations,
halloffame=hof, verbose=False)
return dict([(k, v)
for k, v in zip(self.start.keys(), hof[0])]), None
# CMA_ES solver requires deap > 1.0.1
# http://deap.readthedocs.org/en/latest/examples/cmaes.html
if _deap_available and _numpy_available:
CMA_ES = register_solver('cma-es', 'covariance matrix adaptation evolutionary strategy',
['CMA-ES: covariance matrix adaptation evolutionary strategy',
' ',
'This method requires the following parameters:',
'- num_generations :: number of generations to use',
'- sigma :: (optional) initial covariance, default 1',
'- Lambda :: (optional) measure of reproducibility',
'- starting point: through kwargs'
' ',
'This method is described in detail in:',
'Hansen and Ostermeier, 2001. Completely Derandomized Self-Adaptation in Evolution Strategies. Evolutionary Computation'
])(CMA_ES)