class Individual:
STANDING_LIMITS = {'min': 1, 'max': 10, 'starting': 5}
def __init__(self, exp_class, exp_args=[]):
self.exp_class = exp_class
self.exp_args = exp_args
self._error = 0
self._standing = None
@property
def error(self):
return self._error
@property
def standing(self):
return self._standing
def increment_standing(self):
self._standing = min(self._standing + 1, self.STANDING_LIMITS['max'])
def decrement_standing(self):
self._standing = max(self._standing - 1, self.STANDING_LIMITS['min'])
def init_experiment(self):
self._error = 0
self.experiment = self.exp_class(*self.exp_args)
self.experiment.initialize()
def generate(self, extra_terminal_set=[], extra_function_set=[], tree_depth=3, tree_function_bias=1):
self._standing = self.STANDING_LIMITS['starting']
self.init_experiment()
self.tree = Tree()
self.tree.create(self.experiment.get_terminal_set() + extra_terminal_set,
self.experiment.get_function_set() + extra_function_set,
function_bias=tree_function_bias, max_depth=tree_depth)
def clone(self):
clone = self.__class__(self.exp_class, self.exp_args)
clone._standing = self._standing
clone.init_experiment()
clone.tree = Tree()
clone.tree.clone(self.tree)
return clone
def mutate(self):
mutant = self.__class__(self.exp_class, self.exp_args)
mutant._standing = self._standing
mutant.init_experiment()
mutant.tree = Tree()
mutant.tree.mutate(self.tree)
return mutant
def reproduce(self, other_individual):
child = self.__class__(self.exp_class, self.exp_args)
child._standing = int(numpy.average([self._standing, other_individual._standing]))
child.init_experiment()
child.tree = Tree()
child.tree.subtree_crossover(self.tree, other_individual.tree)
return child
def get_func(self, function_name):
return self.experiment.function_lookup(function_name)
def evaluate(self):
loop = True
while loop:
self._error += self.experiment.norm_error(self.tree.execute(self))
loop = self.experiment.next()
def evaluate_data(self):
samples = []
loop = True
self.experiment.initialize()
while loop:
actual_value = self.tree.execute(self)
sample = {'value': actual_value, 'error': self.experiment.norm_error(actual_value)}
if self.experiment.index() != None:
sample['index'] = self.experiment.index()
samples.append(sample)
loop = self.experiment.next()
return samples
def simplify(self):
self.tree.simplify(self)
