def GTreeMutatorRealGaussian(genome, **args):
""" A gaussian mutator for GTree of Real numbers
Accepts the *rangemin* and *rangemax* genome parameters, both optional. Also
accepts the parameter *gauss_mu* and the *gauss_sigma* which respectively
represents the mean and the std. dev. of the random distribution.
"""
if args["pmut"] <= 0.0:
return 0
elements = len(genome)
mutations = args["pmut"] * elements
mu = genome.getParam("gauss_mu", Consts.CDefG1DListMutRealMU)
sigma = genome.getParam("gauss_sigma", Consts.CDefG1DListMutRealSIGMA)
if mutations < 1.0:
mutations = 0
for i in xrange(len(genome)):
if Util.randomFlipCoin(args["pmut"]):
mutations += 1
rand_node = genome.getRandomNode()
final_value = rand_node.getData() + rand_gauss(mu, sigma)
final_value = min(final_value, genome.getParam("rangemax", Consts.CDefRangeMax))
final_value = max(final_value, genome.getParam("rangemin", Consts.CDefRangeMin))
rand_node.setData(final_value)
else:
for it in xrange(int(round(mutations))):
rand_node = genome.getRandomNode()
final_value = rand_node.getData() + rand_gauss(mu, sigma)
final_value = min(final_value, genome.getParam("rangemax", Consts.CDefRangeMax))
final_value = max(final_value, genome.getParam("rangemin", Consts.CDefRangeMin))
rand_node.setData(final_value)
return int(mutations)
def GTreeMutatorRealGaussian(genome, **args):
""" A gaussian mutator for GTree of Real numbers
Accepts the *rangemin* and *rangemax* genome parameters, both optional. Also
accepts the parameter *gauss_mu* and the *gauss_sigma* which respectively
represents the mean and the std. dev. of the random distribution.
"""
if args["pmut"] <= 0.0:
return 0
elements = len(genome)
mutations = args["pmut"] * elements
mu = genome.getParam("gauss_mu", Consts.CDefG1DListMutRealMU)
sigma = genome.getParam("gauss_sigma", Consts.CDefG1DListMutRealSIGMA)
if mutations < 1.0:
mutations = 0
for i in range(len(genome)):
if Util.randomFlipCoin(args["pmut"]):
mutations += 1
rand_node = genome.getRandomNode()
final_value = rand_node.getData() + rand_gauss(mu, sigma)
final_value = min(final_value, genome.getParam("rangemax", Consts.CDefRangeMax))
final_value = max(final_value, genome.getParam("rangemin", Consts.CDefRangeMin))
rand_node.setData(final_value)
else:
for it in range(int(round(mutations))):
rand_node = genome.getRandomNode()
final_value = rand_node.getData() + rand_gauss(mu, sigma)
final_value = min(final_value, genome.getParam("rangemax", Consts.CDefRangeMax))
final_value = max(final_value, genome.getParam("rangemin", Consts.CDefRangeMin))
rand_node.setData(final_value)
return int(mutations)
def G2DListMutatorRealGaussian(genome, **args):
""" A gaussian mutator for G2DList of Real
Accepts the *rangemin* and *rangemax* genome parameters, both optional. Also
accepts the parameter *gauss_mu* and the *gauss_sigma* which respectively
represents the mean and the std. dev. of the random distribution.
"""
from . import Consts
if args["pmut"] <= 0.0:
return 0
height, width = genome.getSize()
elements = height * width
mutations = args["pmut"] * elements
mu = genome.getParam("gauss_mu")
sigma = genome.getParam("gauss_sigma")
if mu is None:
mu = Consts.CDefG2DListMutRealMU
if sigma is None:
sigma = Consts.CDefG2DListMutRealSIGMA
if mutations < 1.0:
mutations = 0
for i in range(genome.getHeight()):
for j in range(genome.getWidth()):
if Util.randomFlipCoin(args["pmut"]):
final_value = genome[i][j] + rand_gauss(mu, sigma)
final_value = min(
final_value,
genome.getParam("rangemax", Consts.CDefRangeMax))
final_value = max(
final_value,
genome.getParam("rangemin", Consts.CDefRangeMin))
genome.setItem(i, j, final_value)
mutations += 1
else:
for it in range(int(round(mutations))):
which_x = rand_randint(0, genome.getWidth() - 1)
which_y = rand_randint(0, genome.getHeight() - 1)
final_value = genome[which_y][which_x] + rand_gauss(mu, sigma)
final_value = min(final_value,
genome.getParam("rangemax", Consts.CDefRangeMax))
final_value = max(final_value,
genome.getParam("rangemin", Consts.CDefRangeMin))
genome.setItem(which_y, which_x, final_value)
return int(mutations)
def G2DListMutatorIntegerGaussianGradient(genome, **args):
""" A gaussian mutator for G2DList of Integers
Accepts the *rangemin* and *rangemax* genome parameters, both optional.
This routine generates a gaussian value with mu=1.0 and std=0.0333 and then
the gene is multiplied by this value. This will cause the gene to drift
no matter how large it is.
"""
from . import Consts
if args["pmut"] <= 0.0:
return 0
height, width = genome.getSize()
elements = height * width
mutations = args["pmut"] * elements
mu = Consts.CDefGaussianGradientMU
sigma = Consts.CDefGaussianGradientSIGMA
if mutations < 1.0:
mutations = 0
for i in range(genome.getHeight()):
for j in range(genome.getWidth()):
if Util.randomFlipCoin(args["pmut"]):
final_value = int(genome[i][j] *
abs(rand_gauss(mu, sigma)))
final_value = min(
final_value,
genome.getParam("rangemax", Consts.CDefRangeMax))
final_value = max(
final_value,
genome.getParam("rangemin", Consts.CDefRangeMin))
genome.setItem(i, j, final_value)
mutations += 1
else:
for it in range(int(round(mutations))):
which_x = rand_randint(0, genome.getWidth() - 1)
which_y = rand_randint(0, genome.getHeight() - 1)
final_value = int(genome[which_y][which_x] *
abs(rand_gauss(mu, sigma)))
final_value = min(final_value,
genome.getParam("rangemax", Consts.CDefRangeMax))
final_value = max(final_value,
genome.getParam("rangemin", Consts.CDefRangeMin))
genome.setItem(which_y, which_x, final_value)
return int(mutations)
def G2DListMutatorRealGaussian(genome, **args):
""" A gaussian mutator for G2DList of Real
Accepts the *rangemin* and *rangemax* genome parameters, both optional. Also
accepts the parameter *gauss_mu* and the *gauss_sigma* which respectively
represents the mean and the std. dev. of the random distribution.
"""
if args["pmut"] <= 0.0:
return 0
height, width = genome.getSize()
elements = height * width
mutations = args["pmut"] * elements
mu = genome.getParam("gauss_mu")
sigma = genome.getParam("gauss_sigma")
if mu is None:
mu = Consts.CDefG2DListMutRealMU
if sigma is None:
sigma = Consts.CDefG2DListMutRealSIGMA
if mutations < 1.0:
mutations = 0
for i in xrange(genome.getHeight()):
for j in xrange(genome.getWidth()):
if Util.randomFlipCoin(args["pmut"]):
final_value = genome[i][j] + rand_gauss(mu, sigma)
final_value = min(final_value, genome.getParam(
"rangemax", Consts.CDefRangeMax))
final_value = max(final_value, genome.getParam(
"rangemin", Consts.CDefRangeMin))
genome.setItem(i, j, final_value)
mutations += 1
else:
for it in xrange(int(round(mutations))):
which_x = rand_randint(0, genome.getWidth() - 1)
which_y = rand_randint(0, genome.getHeight() - 1)
final_value = genome[which_y][which_x] + rand_gauss(mu, sigma)
final_value = min(final_value, genome.getParam(
"rangemax", Consts.CDefRangeMax))
final_value = max(final_value, genome.getParam(
"rangemin", Consts.CDefRangeMin))
genome.setItem(which_y, which_x, final_value)
return int(mutations)
def G1DListMutatorRealGaussianGradient(genome, **args):
""" The mutator of G1DList, Gaussian Gradient Mutator
Accepts the *rangemin* and *rangemax* genome parameters, both optional. The
random distribution is set with mu=1.0 and std=0.0333
The difference between this routine and the normal Gaussian Real is that the
other function generates a gaussian value and adds it to the value. If the
mu is 0, and the std is 1, a typical value could be 1.8 or -0.5. These small
values are fine if your range is 0-10, but if your range is much larger, like
0-100,000, a relative gradient makes sense.
This routine generates a gaussian value with mu=1.0 and std=0.0333 and then
the gene is multiplied by this value. This will cause the gene to drift
no matter how large it is.
"""
from . import Consts
if args["pmut"] <= 0.0:
return 0
listSize = len(genome)
mutations = args["pmut"] * (listSize)
mu = Consts.CDefGaussianGradientMU
sigma = Consts.CDefGaussianGradientSIGMA
if mutations < 1.0:
mutations = 0
for it in range(listSize):
if Util.randomFlipCoin(args["pmut"]):
final_value = genome[it] * abs(rand_gauss(mu, sigma))
final_value = min(
final_value,
genome.getParam("rangemax", Consts.CDefRangeMax))
final_value = max(
final_value,
genome.getParam("rangemin", Consts.CDefRangeMin))
genome[it] = final_value
mutations += 1
else:
for it in range(int(round(mutations))):
which_gene = rand_randint(0, listSize - 1)
final_value = genome[which_gene] * abs(rand_gauss(mu, sigma))
final_value = min(final_value,
genome.getParam("rangemax", Consts.CDefRangeMax))
final_value = max(final_value,
genome.getParam("rangemin", Consts.CDefRangeMin))
genome[which_gene] = final_value
return int(mutations)
def G2DListMutatorRealGaussianGradient(genome, **args):
""" A gaussian gradient mutator for G2DList of Real
Accepts the *rangemin* and *rangemax* genome parameters, both optional.
The difference is that this multiplies the gene by gauss(1.0, 0.0333), allowing
for a smooth gradient drift about the value.
"""
from . import Consts
if args["pmut"] <= 0.0:
return 0
height, width = genome.getSize()
elements = height * width
mutations = args["pmut"] * elements
mu = Consts.CDefGaussianGradientMU
sigma = Consts.CDefGaussianGradientSIGMA
if mutations < 1.0:
mutations = 0
for i in range(genome.getHeight()):
for j in range(genome.getWidth()):
if Util.randomFlipCoin(args["pmut"]):
final_value = genome[i][j] * abs(rand_gauss(mu, sigma))
final_value = min(
final_value,
genome.getParam("rangemax", Consts.CDefRangeMax))
final_value = max(
final_value,
genome.getParam("rangemin", Consts.CDefRangeMin))
genome.setItem(i, j, final_value)
mutations += 1
else:
for it in range(int(round(mutations))):
which_x = rand_randint(0, genome.getWidth() - 1)
which_y = rand_randint(0, genome.getHeight() - 1)
final_value = genome[which_y][which_x] * abs(rand_gauss(mu, sigma))
final_value = min(final_value,
genome.getParam("rangemax", Consts.CDefRangeMax))
final_value = max(final_value,
genome.getParam("rangemin", Consts.CDefRangeMin))
genome.setItem(which_y, which_x, final_value)
return int(mutations)
def G2DListMutatorIntegerGaussianGradient(genome, **args):
""" A gaussian mutator for G2DList of Integers
Accepts the *rangemin* and *rangemax* genome parameters, both optional.
This routine generates a gaussian value with mu=1.0 and std=0.0333 and then
the gene is multiplied by this value. This will cause the gene to drift
no matter how large it is.
"""
if args["pmut"] <= 0.0:
return 0
height, width = genome.getSize()
elements = height * width
mutations = args["pmut"] * elements
mu = Consts.CDefGaussianGradientMU
sigma = Consts.CDefGaussianGradientSIGMA
if mutations < 1.0:
mutations = 0
for i in xrange(genome.getHeight()):
for j in xrange(genome.getWidth()):
if Util.randomFlipCoin(args["pmut"]):
final_value = int(
genome[i][j] * abs(rand_gauss(mu, sigma)))
final_value = min(final_value, genome.getParam(
"rangemax", Consts.CDefRangeMax))
final_value = max(final_value, genome.getParam(
"rangemin", Consts.CDefRangeMin))
genome.setItem(i, j, final_value)
mutations += 1
else:
for it in xrange(int(round(mutations))):
which_x = rand_randint(0, genome.getWidth() - 1)
which_y = rand_randint(0, genome.getHeight() - 1)
final_value = int(genome[which_y][which_x]
* abs(rand_gauss(mu, sigma)))
final_value = min(final_value, genome.getParam(
"rangemax", Consts.CDefRangeMax))
final_value = max(final_value, genome.getParam(
"rangemin", Consts.CDefRangeMin))
genome.setItem(which_y, which_x, final_value)
return int(mutations)
def G1DListMutatorRealGaussian(genome, **args):
""" The mutator of G1DList, Gaussian Mutator
Accepts the *rangemin* and *rangemax* genome parameters, both optional. Also
accepts the parameter *gauss_mu* and the *gauss_sigma* which respectively
represents the mean and the std. dev. of the random distribution.
"""
from . import Consts
if args["pmut"] <= 0.0:
return 0
listSize = len(genome)
mutations = args["pmut"] * (listSize)
mu = genome.getParam("gauss_mu")
sigma = genome.getParam("gauss_sigma")
if mu is None:
mu = Consts.CDefG1DListMutRealMU
if sigma is None:
sigma = Consts.CDefG1DListMutRealSIGMA
if mutations < 1.0:
mutations = 0
for it in range(listSize):
if Util.randomFlipCoin(args["pmut"]):
final_value = genome[it] + rand_gauss(mu, sigma)
final_value = min(
final_value,
genome.getParam("rangemax", Consts.CDefRangeMax))
final_value = max(
final_value,
genome.getParam("rangemin", Consts.CDefRangeMin))
genome[it] = final_value
mutations += 1
else:
for it in range(int(round(mutations))):
which_gene = rand_randint(0, listSize - 1)
final_value = genome[which_gene] + rand_gauss(mu, sigma)
final_value = min(final_value,
genome.getParam("rangemax", Consts.CDefRangeMax))
final_value = max(final_value,
genome.getParam("rangemin", Consts.CDefRangeMin))
genome[which_gene] = final_value
return int(mutations)
def G1DListMutatorRealGaussianGradient(genome, **args):
""" The mutator of G1DList, Gaussian Gradient Mutator
Accepts the *rangemin* and *rangemax* genome parameters, both optional. The
random distribution is set with mu=1.0 and std=0.0333
The difference between this routine and the normal Gaussian Real is that the
other function generates a gaussian value and adds it to the value. If the
mu is 0, and the std is 1, a typical value could be 1.8 or -0.5. These small
values are fine if your range is 0-10, but if your range is much larger, like
0-100,000, a relative gradient makes sense.
This routine generates a gaussian value with mu=1.0 and std=0.0333 and then
the gene is multiplied by this value. This will cause the gene to drift
no matter how large it is.
"""
if args["pmut"] <= 0.0:
return 0
listSize = len(genome)
mutations = args["pmut"] * (listSize)
mu = Consts.CDefGaussianGradientMU
sigma = Consts.CDefGaussianGradientSIGMA
if mutations < 1.0:
mutations = 0
for it in xrange(listSize):
if Util.randomFlipCoin(args["pmut"]):
final_value = genome[it] * abs(rand_gauss(mu, sigma))
final_value = min(final_value, genome.getParam(
"rangemax", Consts.CDefRangeMax))
final_value = max(final_value, genome.getParam(
"rangemin", Consts.CDefRangeMin))
genome[it] = final_value
mutations += 1
else:
for it in xrange(int(round(mutations))):
which_gene = rand_randint(0, listSize - 1)
final_value = genome[which_gene] * abs(rand_gauss(mu, sigma))
final_value = min(final_value, genome.getParam(
"rangemax", Consts.CDefRangeMax))
final_value = max(final_value, genome.getParam(
"rangemin", Consts.CDefRangeMin))
genome[which_gene] = final_value
return int(mutations)
def G1DListMutatorAlleleGaussian(genome, **arguments):
"""An allele-based mutator based on G1DListMutatorRealGaussian.
Accepts the parameter *gauss_mu* and the *gauss_sigma* which
respectively represents the mean and the std. dev. of the random
distribution.
"""
from . import Consts
if arguments["pmut"] <= 0.0:
return 0
listSize = len(genome)
mutations = arguments["pmut"] * listSize
mu = genome.getParam("gauss_mu")
sigma = genome.getParam("gauss_sigma")
if mu is None:
mu = Consts.CDefG1DListMutRealMU
if sigma is None:
sigma = Consts.CDefG1DListMutRealSIGMA
allele = genome.getParam("allele", None)
if allele is None:
Util.raiseException(
"to use this mutator, you must specify the 'allele' parameter",
TypeError)
if mutations < 1.0:
mutations = 0
for it in range(listSize):
if Util.randomFlipCoin(arguments["pmut"]):
final_value = genome[it] + rand_gauss(mu, sigma)
assert len(allele[it].beginEnd
) == 1, "only single ranges are supported"
rangemin, rangemax = allele[it].beginEnd[0]
final_value = min(final_value, rangemax)
final_value = max(final_value, rangemin)
genome[it] = final_value
mutations += 1
else:
for it in range(int(round(mutations))):
which_gene = rand_randint(0, listSize - 1)
final_value = genome[which_gene] + rand_gauss(mu, sigma)
assert len(allele[which_gene].beginEnd
) == 1, "only single ranges are supported"
rangemin, rangemax = allele[which_gene].beginEnd[0]
final_value = min(final_value, rangemax)
final_value = max(final_value, rangemin)
genome[which_gene] = final_value
return int(mutations)
def G1DListMutatorIntegerGaussianGradient(genome, **args):
""" A gaussian mutator for G1DList of Integers
Accepts the *rangemin* and *rangemax* genome parameters, both optional. The
random distribution is set with mu=1.0 and std=0.0333
Same as IntegerGaussian, except that this uses relative gradient rather than
absolute gaussian. A value is randomly generated about gauss(mu=1, sigma=.0333)
and multiplied by the gene to drift it up or down (depending on what side of
1 the random value falls on) and cast to integer
"""
from . import Consts
if args["pmut"] <= 0.0:
return 0
listSize = len(genome)
mutations = args["pmut"] * (listSize)
mu = Consts.CDefGaussianGradientMU
sigma = Consts.CDefGaussianGradientSIGMA
if mutations < 1.0:
mutations = 0
for it in range(listSize):
if Util.randomFlipCoin(args["pmut"]):
final_value = int(genome[it] * abs(rand_gauss(mu, sigma)))
final_value = min(
final_value,
genome.getParam("rangemax", Consts.CDefRangeMax))
final_value = max(
final_value,
genome.getParam("rangemin", Consts.CDefRangeMin))
genome[it] = final_value
mutations += 1
else:
for it in range(int(round(mutations))):
which_gene = rand_randint(0, listSize - 1)
final_value = int(genome[which_gene] * abs(rand_gauss(mu, sigma)))
final_value = min(final_value,
genome.getParam("rangemax", Consts.CDefRangeMax))
final_value = max(final_value,
genome.getParam("rangemin", Consts.CDefRangeMin))
genome[which_gene] = final_value
return int(mutations)
def G1DListMutatorRealGaussian(genome, **args):
""" The mutator of G1DList, Gaussian Mutator
Accepts the *rangemin* and *rangemax* genome parameters, both optional. Also
accepts the parameter *gauss_mu* and the *gauss_sigma* which respectively
represents the mean and the std. dev. of the random distribution.
"""
# args["pmut"]=ga.setMutationRate(0.98)
if args["pmut"] <= 0.0: return 0
listSize = len(genome)
mutations = args["pmut"] * (listSize)
mu = genome.getParam("gauss_mu")
sigma = genome.getParam("gauss_sigma")
length=genome.getParam("length",2)
range_list=genome.getParam("range_list",None)
if mu is None:
mu = Consts.CDefG1DListMutRealMU
if sigma is None:
sigma = Consts.CDefG1DListMutRealSIGMA
if mutations < 1.0:
mutations = 0
for it in xrange(listSize):
if Util.randomFlipCoin(args["pmut"]):
final_value = round(genome[it] + rand_gauss(mu, sigma),length)
if range_list is None:
final_value = min(final_value, genome.getParam("rangemax", Consts.CDefRangeMax))
final_value = max(final_value, genome.getParam("rangemin", Consts.CDefRangeMin))
else:
final_value = min(final_value, range_list[it][1])
final_value = max(final_value, range_list[it][0])
genome[it] = final_value
mutations += 1
else:
for it in xrange(int(round(mutations))):
which_gene = rand_randint(0, listSize-1)
final_value = round(genome[which_gene] + rand_gauss(mu, sigma),length)
range_list=genome.getParam("range_list",None)
if range_list is None:
final_value = min(final_value, genome.getParam("rangemax", Consts.CDefRangeMax))
final_value = max(final_value, genome.getParam("rangemin", Consts.CDefRangeMin))
else:
final_value = min(final_value, range_list[which_gene][1])
final_value = max(final_value, range_list[which_gene][0])
genome[which_gene] = final_value
return int(mutations)
def G2DListMutatorRealGaussianGradient(genome, **args):
""" A gaussian gradient mutator for G2DList of Real
Accepts the *rangemin* and *rangemax* genome parameters, both optional.
The difference is that this multiplies the gene by gauss(1.0, 0.0333), allowing
for a smooth gradient drift about the value.
"""
if args["pmut"] <= 0.0:
return 0
height, width = genome.getSize()
elements = height * width
mutations = args["pmut"] * elements
mu = Consts.CDefGaussianGradientMU
sigma = Consts.CDefGaussianGradientSIGMA
if mutations < 1.0:
mutations = 0
for i in xrange(genome.getHeight()):
for j in xrange(genome.getWidth()):
if Util.randomFlipCoin(args["pmut"]):
final_value = genome[i][j] * abs(rand_gauss(mu, sigma))
final_value = min(final_value, genome.getParam(
"rangemax", Consts.CDefRangeMax))
final_value = max(final_value, genome.getParam(
"rangemin", Consts.CDefRangeMin))
genome.setItem(i, j, final_value)
mutations += 1
else:
for it in xrange(int(round(mutations))):
which_x = rand_randint(0, genome.getWidth() - 1)
which_y = rand_randint(0, genome.getHeight() - 1)
final_value = genome[which_y][which_x] * abs(rand_gauss(mu, sigma))
final_value = min(final_value, genome.getParam(
"rangemax", Consts.CDefRangeMax))
final_value = max(final_value, genome.getParam(
"rangemin", Consts.CDefRangeMin))
genome.setItem(which_y, which_x, final_value)
return int(mutations)
def G1DListMutatorIntegerGaussian(genome, **args):
""" A gaussian mutator for G1DList of Integers
Accepts the *rangemin* and *rangemax* genome parameters, both optional. Also
accepts the parameter *gauss_mu* and the *gauss_sigma* which respectively
represents the mean and the std. dev. of the random distribution.
"""
if args["pmut"] <= 0.0:
return 0
listSize = len(genome)
mutations = args["pmut"] * (listSize)
mu = genome.getParam("gauss_mu")
sigma = genome.getParam("gauss_sigma")
if mu is None:
mu = Consts.CDefG1DListMutIntMU
if sigma is None:
sigma = Consts.CDefG1DListMutIntSIGMA
if mutations < 1.0:
mutations = 0
for it in xrange(listSize):
if Util.randomFlipCoin(args["pmut"]):
final_value = genome[it] + int(rand_gauss(mu, sigma))
final_value = min(final_value, genome.getParam(
"rangemax", Consts.CDefRangeMax))
final_value = max(final_value, genome.getParam(
"rangemin", Consts.CDefRangeMin))
genome[it] = final_value
mutations += 1
else:
for it in xrange(int(round(mutations))):
which_gene = rand_randint(0, listSize - 1)
final_value = genome[which_gene] + int(rand_gauss(mu, sigma))
final_value = min(final_value, genome.getParam(
"rangemax", Consts.CDefRangeMax))
final_value = max(final_value, genome.getParam(
"rangemin", Consts.CDefRangeMin))
genome[which_gene] = final_value
return int(mutations)
def G1DConnBiasedMutateWeights(genome, **args):
## args['mutation_conns'] takes the list of indices of connections to be mutated
if not args['mutation_conns']:
mutation_conn_indices=rand_sample(numpy.array(len(genome)-1),int(args['pmut']*len(genome)))
# print "Indices of mutation weights:"
# print mutation_conn_indices
else:
mutation_conn_indices = args['mutation_conns']
# print "Indices of mutation weights:"
# print mutation_conn_indices
mu = genome.getParam("gauss_mu")
sigma = genome.getParam("gauss_sigma")
new_mutation_list=[]
for it in mutation_conn_indices:
final_value = genome['weight'][it] + rand_gauss(mu, sigma)
#final_value = min(final_value, genome.getParam("rangemax", Consts.CdefRangeMax))
#final_value = max(final_value, genome.getParam("rangemin", Consts.CdefRangeMin))
new_mutation_list.append(final_value)
numpy.put(genome['weight'],[mutation_conn_indices],[new_mutation_list])
return genome
def G1DListMutatorAlleleGaussian(genome, **arguments):
"""An allele-based mutator based on G1DListMutatorRealGaussian.
Accepts the parameter *gauss_mu* and the *gauss_sigma* which
respectively represents the mean and the std. dev. of the random
distribution.
"""
if arguments["pmut"] <= 0.0:
return 0
listSize = len(genome)
mutations = arguments["pmut"] * listSize
mu = genome.getParam("gauss_mu")
sigma = genome.getParam("gauss_sigma")
if mu is None:
mu = Consts.CDefG1DListMutRealMU
if sigma is None:
sigma = Consts.CDefG1DListMutRealSIGMA
allele = genome.getParam("allele", None)
if allele is None:
Util.raiseException(
"to use this mutator, you must specify the 'allele' parameter", TypeError)
if mutations < 1.0:
mutations = 0
for it in xrange(listSize):
if Util.randomFlipCoin(arguments["pmut"]):
final_value = genome[it] + rand_gauss(mu, sigma)
assert len(
allele[it].beginEnd) == 1, "only single ranges are supported"
rangemin, rangemax = allele[it].beginEnd[0]
final_value = min(final_value, rangemax)
final_value = max(final_value, rangemin)
genome[it] = final_value
mutations += 1
else:
for it in xrange(int(round(mutations))):
which_gene = rand_randint(0, listSize - 1)
final_value = genome[which_gene] + rand_gauss(mu, sigma)
assert len(
allele[which_gene].beginEnd) == 1, "only single ranges are supported"
rangemin, rangemax = allele[which_gene].beginEnd[0]
final_value = min(final_value, rangemax)
final_value = max(final_value, rangemin)
genome[which_gene] = final_value
return int(mutations)
def G1DListMutatorIntegerGaussianGradient(genome, **args):
""" A gaussian mutator for G1DList of Integers
Accepts the *rangemin* and *rangemax* genome parameters, both optional. The
random distribution is set with mu=1.0 and std=0.0333
Same as IntegerGaussian, except that this uses relative gradient rather than
absolute gaussian. A value is randomly generated about gauss(mu=1, sigma=.0333)
and multiplied by the gene to drift it up or down (depending on what side of
1 the random value falls on) and cast to integer
"""
if args["pmut"] <= 0.0:
return 0
listSize = len(genome)
mutations = args["pmut"] * (listSize)
mu = Consts.CDefGaussianGradientMU
sigma = Consts.CDefGaussianGradientSIGMA
if mutations < 1.0:
mutations = 0
for it in xrange(listSize):
if Util.randomFlipCoin(args["pmut"]):
final_value = int(genome[it] * abs(rand_gauss(mu, sigma)))
final_value = min(final_value, genome.getParam(
"rangemax", Consts.CDefRangeMax))
final_value = max(final_value, genome.getParam(
"rangemin", Consts.CDefRangeMin))
genome[it] = final_value
mutations += 1
else:
for it in xrange(int(round(mutations))):
which_gene = rand_randint(0, listSize - 1)
final_value = int(genome[which_gene] * abs(rand_gauss(mu, sigma)))
final_value = min(final_value, genome.getParam(
"rangemax", Consts.CDefRangeMax))
final_value = max(final_value, genome.getParam(
"rangemin", Consts.CDefRangeMin))
genome[which_gene] = final_value
return int(mutations)
def G1DConnInitializatorGaussian(genome, **args):
""" Real initialization function of G1DConnections
This initializator accepts the *rangemin* and *rangemax* genome parameters.
"""
range_min = genome.getParam("rangemin", -50)
range_max = genome.getParam("rangemax", 50)
mu = genome.getParam("gauss_mu", 0)
sigma = genome.getParam("gauss_sigma", 1)
# print "mu=", mu, "sigma=", sigma
layers = genome.getParam("layers", [2,4,1])
bias = genome.getParam("bias", [1,1,0])
ndLayers=np.array(layers)
ndBias=np.array(bias)
if genome.genomeList == []:
totalConnections = int(np.sum([(layers[i]+1)*layers[i+1] for i in range(len(layers)-1)]))
genomeList1=[]
for ly in range(len(layers)-1):
for toNeurode in range(layers[ly+1]):
toNeurode += int(ndLayers[0:ly+1].sum()+ndBias[0:ly+1].sum())
for fromNeurode in range(layers[ly]+1):
fromNeurode +=int(ndLayers[0:ly].sum()+ndBias[0:ly].sum())
weight = min(range_max, rand_gauss(mu, sigma))
weight = max(range_min, weight)
genomeList1.append((fromNeurode, toNeurode, weight))
genome.genomeList = np.array(genomeList1, dtype=[('from','i'),('to','i'),('weight','f')])
else:
totalConnections = len(genome.genomeList)
if totalConnections != genome.genomeSize:
print "%% Warning: genome.genomeSize and the length of genome.genomeList mismatch!"
for conn in genome.genomeList:
weight = min(range_max, rand_gauss(mu, sigma))
weight = max(range_min, weight)
conn[2] = weight
def G1DConnUnbiasedMutateWeights(genome, **args):
## args['mutation_conns'] takes the list of indices of connections to be mutated
if not args.get('mutation_conns',0):
mutation_conn_indices=rand_sample(numpy.arange(len(genome)),int(args['pmut']*len(genome)))
# print "Indices of mutation weights:"
# print mutation_conn_indices
else:
mutation_conn_indices = args['mutation_conns']
# print "Indices of mutation weights:"
# print mutation_conn_indices
mu = genome.getParam("gauss_mu",0)
sigma = genome.getParam("gauss_sigma",1)
#new_mutation_list=[]
for it in mutation_conn_indices:
final_value=rand_gauss(mu, sigma)
genome['weight'][it] = final_value
# new_mutation_list.append(rand_uniform(genome.getParam("rangemin", Consts.CDefRangeMin),genome.getParam("rangemax", Consts.CDefRangeMax)))
mutations = len(mutation_conn_indices)
#numpy.put(genome['weight'],mutation_conn_indices,new_mutation_list)
return int(mutations)
def G1DListMutatorIntegerGaussian(genome, **args):
""" A gaussian mutator for G1DList of Integers
Accepts the *rangemin* and *rangemax* genome parameters, both optional. Also
accepts the parameter *gauss_mu* and the *gauss_sigma* which respectively
represents the mean and the std. dev. of the random distribution.
"""
if args["pmut"] <= 0.0: return 0
listSize = len(genome)
mutations = args["pmut"] * (listSize)
range_list=genome.getParam("range_list",None)
mu = genome.getParam("gauss_mu")
sigma = genome.getParam("gauss_sigma")
if mu is None:
mu = Consts.CDefG1DListMutIntMU
if sigma is None:
sigma = Consts.CDefG1DListMutIntSIGMA
if mutations < 1.0:
mutations = 0
for it in xrange(listSize):
if Util.randomFlipCoin(args["pmut"]):
final_value = genome[it] + int(rand_gauss(mu, sigma))
if range_list is None:
final_value = min(final_value, genome.getParam("rangemax", Consts.CDefRangeMax))
final_value = max(final_value, genome.getParam("rangemin", Consts.CDefRangeMin))
else:
final_value = min(final_value, range_list[it])
final_value = max(final_value, 0)
genome[it] = final_value
mutations += 1
else:
for it in xrange(int(round(mutations))):
which_gene = rand_randint(0, listSize-1)
final_value = genome[which_gene] + int(rand_gauss(mu, sigma))
if range_list is None:
final_value = min(final_value, genome.getParam("rangemax", Consts.CDefRangeMax))
final_value = max(final_value, genome.getParam("rangemin", Consts.CDefRangeMin))
else:
final_value = min(final_value,range_list[which_gene])
final_value = max(final_value, 0)
genome[which_gene] = final_value
index=genome.getParam("index1",None)
margin=genome.getParam("margin",None)
if index!=None or margin!=None:
while(1):
sum=0
for i in index:
sum=sum+genome[i]*margin[i]
if sum<400000:
break
for j in index:
if genome[j]>0:
genome[j]=genome[j]-1
sum=sum-margin[j]
#print genome[0:]
if sum<400000:
break
#print genome[0:]
return int(mutations)
def G2DListMutatorRealGaussianSpline(genome, **args):
""" A gaussian mutator for G2DList of Real with UnivariateSpline for smoothing
Accepts the *rangemin* and *rangemax* genome parameters, both optional. Also
accepts the parameter *gauss_mu* and the *gauss_sigma* which respectively
represents the mean and the std. dev. of the random distribution. """
dtgenome = DTGenomeFunctions()
pmut = args["pmut"]
if pmut <= 0.0: return 0
width = dtgenome.getIndividualFrameLength(genome)
height = dtgenome.getIndividualLength(genome)
elements = height * width
mutations = pmut * elements
mu = genome.getParam("gauss_mu")
sigma = genome.getParam("gauss_sigma")
prop = DTIndividualPropertyVanillaEvolutive()
robotConfig = LoadRobotConfiguration()
robotJoints = robotConfig.getJointsName()
validIDs = []
jointIndex = 0
for joint in robotJoints:
if not prop.avoidJoint(joint):
validIDs.append(jointIndex)
jointIndex += 1
if mu is None:
mu = CDefG2DListMutRealMU
if sigma is None:
sigma = CDefG2DListMutRealSIGMA
if mutations < 1.0:
mutations = 0
for i in xrange(genome.getHeight()):
for j in xrange(genome.getWidth()):
if genome[i][j] != sysConstants.JOINT_SENTINEL and j in validIDs and randomFlipCoin(pmut):
offset = rand_gauss(mu, sigma)
print "OFFSET: ", offset, "joint: ", j, "frame: ", i
final_value = genome[i][j] + offset
final_value = min(final_value, genome.getParam("rangemax", CDefRangeMax))
final_value = max(final_value, genome.getParam("rangemin", CDefRangeMin))
genome.setItem(i, j, final_value)
dtgenome.smooth(genome, j, i, offset)
##pca.poseInterpolationWithPCA(chromosomeToLucyGeneticMatrix(genome), j)
mutations += 1
else:
for it in xrange(int(round(mutations))):
foundFrameDifferentThanSentinel = False
validIDFound = False
while not foundFrameDifferentThanSentinel and not validIDFound:
which_x = rand_randint(0, genome.getWidth() - 1) # joint to mutate
if which_x in validIDs: #only mutate not avoided ids
validIDFound = True
which_y = rand_randint(0, genome.getHeight() - 1) # pose to mutate
if genome[which_y][which_x] != sysConstants.JOINT_SENTINEL:
foundFrameDifferentThanSentinel = True
offset = rand_gauss(mu, sigma)
final_value = genome[which_y][which_x] + offset
#print "OFFSET: ", offset, "joint: ", which_x, "frame: ", which_y
# to be sure that the value is less than the rangemax and more than the rangemin value
final_value = min(final_value, genome.getParam("rangemax", CDefRangeMax))
final_value = max(final_value, genome.getParam("rangemin", CDefRangeMin))
# valueBeforeMutation = genome[which_y][which_x]
genome.setItem(which_y, which_x, final_value)
dtgenome.smooth(genome, which_x, which_y)
##pca.poseInterpolationWithPCA(chromosomeToLucyGeneticMatrix(genome), which_x)
return int(mutations)
