Problem, GP with Island model.

[COMZ]

I merged two example; gp_symbreg.py and ga_symbreg_multidemic.py. But,
it did not work. How can i solve it?


import array
import random

from deap import algorithms
from deap import base
from deap import creator
from deap import tools

import operator
import math
from deap import gp

def safeDiv(left, right):
try:
return left / right
except ZeroDivisionError:
return 0

pset = gp.PrimitiveSet("MAIN", 1)
pset.addPrimitive(operator.add, 2)
pset.addPrimitive(operator.sub, 2)
pset.addPrimitive(operator.mul, 2)
pset.addPrimitive(safeDiv, 2)
pset.addPrimitive(operator.neg, 1)
pset.addPrimitive(math.cos, 1)
pset.addPrimitive(math.sin, 1)
pset.addEphemeralConstant(lambda: random.randint(-1,1))
pset.renameArguments({"ARG0" : "x"})

creator.create("FitnessMin", base.Fitness, weights=(-1.0,))
creator.create("Individual", gp.PrimitiveTree,
fitness=creator.FitnessMin, pset=pset)

toolbox = base.Toolbox()

toolbox.register("expr", gp.genRamped, pset=pset, min_=1, max_=2)
toolbox.register("individual", tools.initIterate, creator.Individual,
toolbox.expr)
toolbox.register("population", tools.initRepeat, list,
toolbox.individual)
toolbox.register("lambdify", gp.lambdify, pset=pset)

def evalSymbReg(individual):
# Transform the tree expression in a callable function
func = toolbox.lambdify(expr=individual)
# Evaluate the sum of squared difference between the expression
# and the real function : x**4 + x**3 + x**2 + x
values = (x/10. for x in xrange(-10,10))
diff_func = lambda x: (func(x)-(x**4 + x**3 + x**2 + x))**2
diff = sum(map(diff_func, values))
return diff,

toolbox.register("evaluate", evalSymbReg)
toolbox.register("select", tools.selTournament, tournsize=3)
toolbox.register("mate", gp.cxUniformOnePoint)
toolbox.register("expr_mut", gp.genFull, min_=0, max_=2)
toolbox.register('mutate', gp.mutUniform, expr=toolbox.expr_mut)
toolbox.register("migrate", tools.migRing, k=5,
selection=tools.selBest, replacement=tools.selRandom)

def main():
random.seed(64)

NBR_DEMES = 3
MU = 30
NGEN = 40
CXPB = 0.5
MUTPB = 0.2
MIG_RATE = 5

demes = [toolbox.population(n=MU) for _ in xrange(NBR_DEMES)]
hof = tools.HallOfFame(1)
stats = tools.Statistics(lambda ind: ind.fitness.values, 4)
stats.register("avg", tools.mean)
stats.register("std", tools.std)
stats.register("min", min)
stats.register("max", max)

logger = tools.EvolutionLogger(["gen", "evals"] +
stats.functions.keys())
logger.logHeader()

for idx, deme in enumerate(demes):
for ind in deme:
ind.fitness.values = toolbox.evaluate(ind)
stats.update(deme, idx)
hof.update(deme)
logger.logGeneration(gen="0.%d" % idx, evals=len(deme),
stats=stats, index=idx)

stats.update(demes[0]+demes[1]+demes[2], 3)
logger.logGeneration(gen=0, evals="-", stats=stats, index=3)

gen = 1
while gen <= NGEN and stats.min[3][-1][0] > 0 :#stats.max[3][-1]
[0] < 100.0:
# print stats.max[3][-1][0],stats.min[3][-1][0]
for idx, deme in enumerate(demes):
algorithms.varAnd(deme, toolbox, cxpb=CXPB, mutpb=MUTPB)
#algorithms.eaSimple(deme, toolbox, CXPB, MUTPB, 1, stats,
halloffame=hof)
for ind in deme:
ind.fitness.values = toolbox.evaluate(ind)

stats.update(deme, idx)
hof.update(deme)
logger.logGeneration(gen="%d.%d" % (gen, idx),
evals=len(deme), stats=stats, index=idx)

if gen % MIG_RATE == 0:
toolbox.migrate(demes)
stats.update(demes[0]+demes[1]+demes[2], 3)
logger.logGeneration(gen="%d" % gen, evals="-", stats=stats,
index=3)
gen += 1

print hof
return demes, stats, hof

if __name__ == "__main__":
main()




ERROR MESG:

File "island.py", line 130, in <module>
main()
File "island.py", line 121, in main
toolbox.migrate(demes)
File "/opt/local/Library/Frameworks/Python.framework/Versions/2.7/
lib/python2.7/site-packages/deap/tools.py", line 1786, in migRing
indx = populations[to_deme].index(immigrant)
File "/opt/local/Library/Frameworks/Python.framework/Versions/2.7/
lib/python2.7/site-packages/deap/base.py", line 279, in __eq__
return self.obj == other.obj
AttributeError: 'Individual' object has no attribute 'obj'

[Félix-Antoine Fortin]

Hi,

Following your message, we found two problems, on in deap.base and 

the other and the multidemic example. So first, thank you for your bug

report ;).

Second, to get your code working, you will need to modifiy deap/base.py,

more specifically, you will need to replace the __eq__ of NodeProxy by

the following code :

        def __eq__(self, other):
            if isistance(other, self.__class__):
                return self.obj == other.obj
            else:

                return False

Basically, the problem was that a Tree can be composed of Node, but

also other trees. Since a tree is a list, Python use the __eq__ of list to

determine if two trees are equals. However, the __eq__ operator was 

not conceived to compare a Node with anything else than a Node. 

With the preceding patch, if the other object is not a Node, the function

now returns false.

As to why this error only surfaces when using migration, it is because 

we almost never use check if two trees are equal. However, in migRing,

we use the list.index operation to find the index associated to an Individual

in a list of Individual, and list.index use the __eq__ operator to compare

every Individual of the list with the argument. That concludes the problem

you reported with migRing.

We however found a second problem in the configuration of ga_onemax_multidemic.

It is explicited clearly in the documentation that : 

the replacement strategy must select k 

different individuals. For example, using a traditional tournament for 

replacement strategy will thus give undesirable effects, two individuals 

will most likely try to enter the same slot.

However in our own example, we used selRandom as a replacement strategy,

which contradicts the preceding quotes since this operator can return the same

individual multiple times. Therefore, in your code instead of using selRandom,

you should use random.sample :

toolbox.register("migrate", tools.migRing, k=5, 

selection=tools.selBest, replacement=random.sample) 

Both of this problems have been fixed on the default branch of the Mercurial

repository, and will be part of the next release, 0.8.2, which should come 

out soon.

Thanks again for the bug report.

Félix-Antoine Fortin

[Jongseong KIM]

Hi,

Thank you for answer.

Always, I think DEAP is really awesome.

Jongseong Kim

Ph.d Candidate

Department of Industrial Engineering

Sungkyunkwan Univ.

[Jongseong KIM]

Hi,

I checkout repository. It occur bug.

You missed. 

Correct following,please.

isistance ----> isinstance

Jongseong Kim

Ph.d Candidate

Department of Industrial Engineering

Sungkyunkwan Univ.

On 2012. 6. 7., at 4:06, "Félix-Antoine Fortin" <felix.anto...@gmail.com> wrote:

[Félix-Antoine Fortin]

Oups, sorry about that.

It is now fixed.

Thanks again.

Félix-Antoine Fortin

On Tuesday, June 5, 2012 10:58:12 PM UTC-4, COMZ wrote: