NiaPy.algorithms.algorithm.Individual.evaluate() - Code Metrics - Inspection of "Merge pull request #233 from kb2623/development" - NiaOrg/NiaPy - Measure and Improve Code Quality continuously with Scrutinizer

Completed

Push — master ( cc7745...279fb5 )

by Grega

created 2019-11-12 07:54 UTC

NiaPy.algorithms.algorithm.Individual.evaluate() A

↳ Parent: NiaPy.algorithms.algorithm

Complexity

Conditions

Size

Total Lines	15
Code Lines	3

Duplication

Lines	0
Ratio	0 %

Importance

Changes

Metric	Value
cc	1
eloc	3
nop	3
dl	0
loc	15
rs	10
c	0
b	0
f	0

# encoding=utf8
import logging

from numpy import random as rand, inf, ndarray, asarray, array_equal, argmin, apply_along_axis

from NiaPy.util import FesException, GenException, TimeException, RefException
from NiaPy.util.utility import objects2array

logging.basicConfig()
logger = logging.getLogger('NiaPy.util.utility')
logger.setLevel('INFO')

__all__ = [
	'Algorithm',
	'Individual',
	'defaultIndividualInit',
	'defaultNumPyInit'
]

def defaultNumPyInit(task, NP, rnd=rand, **kwargs):
	r"""Initialize starting population that is represented with `numpy.ndarray` with shape `{NP, task.D}`.

	Args:
		task (Task): Optimization task.
		NP (int): Number of individuals in population.
		rnd (Optional[mtrand.RandomState]): Random number generator.
		kwargs (Dict[str, Any]): Additional arguments.

	Returns:
		Tuple[numpy.ndarray, numpy.ndarray[float]]:
			1. New population with shape `{NP, task.D}`.
			2. New population function/fitness values.
	"""
	pop = task.Lower + rnd.rand(NP, task.D) * task.bRange
	fpop = apply_along_axis(task.eval, 1, pop)
	return pop, fpop

def defaultIndividualInit(task, NP, rnd=rand, itype=None, **kwargs):
	r"""Initialize `NP` individuals of type `itype`.

	Args:
		task (Task): Optimization task.
		NP (int): Number of individuals in population.
		rnd (Optional[mtrand.RandomState]): Random number generator.
		itype (Optional[Individual]): Class of individual in population.
		kwargs (Dict[str, Any]): Additional arguments.

	Returns:
		Tuple[numpy.ndarray[Individual], numpy.ndarray[float]:
			1. Initialized individuals.
			2. Initialized individuals function/fitness values.
	"""
	pop = objects2array([itype(task=task, rnd=rnd, e=True) for _ in range(NP)])
	return pop, asarray([x.f for x in pop])

class Algorithm:
	r"""Class for implementing algorithms.

	Date:
		2018

	Author
		Klemen Berkovič

	License:
		MIT

	Attributes:
		Name (List[str]): List of names for algorithm.
		Rand (mtrand.RandomState): Random generator.
		NP (int): Number of inidividuals in populatin.
		InitPopFunc (Callable[[int, Task, mtrand.RandomState, Dict[str, Any]], Tuple[numpy.ndarray, numpy.ndarray[float]]]): Idividual initialization function.
		itype (Individual): Type of individuals used in population, default value is None for Numpy arrays.
	"""
	Name = ['Algorithm', 'AAA']
	Rand = rand.RandomState(None)

	NP = 50
	InitPopFunc = defaultNumPyInit
	itype = None

	@staticmethod
	def typeParameters():
		r"""Return functions for checking values of parameters.

		Return:
			Dict[str, Callable]:
				* NP (Callable[[int], bool]): Check if number of individuals is :math:`\in [0, \infty]`.
		"""
		return {'NP': lambda x: isinstance(x, int) and x > 0}

	def __init__(self, **kwargs):
		r"""Initialize algorithm and create name for an algorithm.

		Args:
			seed (int): Starting seed for random generator.

		See Also:
			* :func:`NiaPy.algorithms.Algorithm.setParameters`
		"""
		self.Rand, self.exception = rand.RandomState(kwargs.pop('seed', None)), None
		self.setParameters(**kwargs)

	@staticmethod
	def algorithmInfo():
		r"""Get algorithm information.

		Returns:
			str: Bit item.
		"""
		return '''Basic algorithm. No implementation!!!'''

	def setParameters(self, NP=50, InitPopFunc=defaultNumPyInit, itype=None, **kwargs):
		r"""Set the parameters/arguments of the algorithm.

		Args:
			NP (Optional[int]): Number of individuals in population :math:`\in [1, \infty]`.
			InitPopFunc (Optional[Callable[[int, Task, mtrand.RandomState, Dict[str, Any]], Tuple[numpy.ndarray, numpy.ndarray[float]]]]): Type of individuals used by algorithm.
			itype (Optional[Any]): Individual type used in population, default is Numpy array.
			**kwargs (Dict[str, Any]): Additional arguments.

		See Also:
			* :func:`NiaPy.algorithms.defaultNumPyInit`
			* :func:`NiaPy.algorithms.defaultIndividualInit`
		"""
		self.NP, self.InitPopFunc, self.itype = NP, InitPopFunc, itype

	def getParameters(self):
		r"""Get parameters of the algorithm.

		Returns:
			Dict[str, Any]:
			* Parameter name (str): Represents a parameter name
			* Value of parameter (Any): Represents the value of the parameter
		"""
		return {
			'NP': self.NP,
			'InitPopFunc': self.InitPopFunc,
			'itype': self.itype
		}

	def rand(self, D=1):
		r"""Get random distribution of shape D in range from 0 to 1.

		Args:
			D (numpy.ndarray[int]): Shape of returned random distribution.

		Returns:
			Union[numpy.ndarray[float], float]: Random number or numbers :math:`\in [0, 1]`.
		"""
		if isinstance(D, (ndarray, list)): return self.Rand.rand(*D)
		elif D > 1: return self.Rand.rand(D)
		else: return self.Rand.rand()

	def uniform(self, Lower, Upper, D=None):
		r"""Get uniform random distribution of shape D in range from "Lower" to "Upper".

		Args:
			Lower (Iterable[float]): Lower bound.
			Upper (Iterable[float]): Upper bound.
			D (Union[int, Iterable[int]]): Shape of returned uniform random distribution.

		Returns:
			Union[numpy.ndarray[float], float]: Array of numbers :math:`\in [\mathit{Lower}, \mathit{Upper}]`.
		"""
		return self.Rand.uniform(Lower, Upper, D) if D is not None else self.Rand.uniform(Lower, Upper)

	def normal(self, loc, scale, D=None):
		r"""Get normal random distribution of shape D with mean "loc" and standard deviation "scale".

		Args:
			loc (float): Mean of the normal random distribution.
			scale (float): Standard deviation of the normal random distribution.
			D (Union[int, Iterable[int]]): Shape of returned normal random distribution.

		Returns:
			Union[numpy.ndarray[float], float]: Array of numbers.
		"""
		return self.Rand.normal(loc, scale, D) if D is not None else self.Rand.normal(loc, scale)

	def randn(self, D=None):
		r"""Get standard normal distribution of shape D.

		Args:
			D (Union[int, Iterable[int]]): Shape of returned standard normal distribution.

		Returns:
			Union[numpy.ndarray[float], float]: Random generated numbers or one random generated number :math:`\in [0, 1]`.
		"""
		if D is None: return self.Rand.randn()
		elif isinstance(D, int): return self.Rand.randn(D)
		return self.Rand.randn(*D)

	def randint(self, Nmax, D=1, Nmin=0, skip=None):
		r"""Get discrete uniform (integer) random distribution of D shape in range from "Nmin" to "Nmax".

		Args:
			Nmin (int): Lower integer bound.
			Nmax (int): One above upper integer bound.
			D (Union[int, Iterable[int]]): shape of returned discrete uniform random distribution.
			skip (Union[int, Iterable[int], numpy.ndarray[int]]): numbers to skip.

		Returns:
			Union[int, numpy.ndarrayj[int]]: Random generated integer number.
		"""
		r = None
		if isinstance(D, (list, tuple, ndarray)): r = self.Rand.randint(Nmin, Nmax, D)
		elif D > 1: r = self.Rand.randint(Nmin, Nmax, D)
		else: r = self.Rand.randint(Nmin, Nmax)
		return r if skip is None or r not in skip else self.randint(Nmax, D, Nmin, skip)

	def getBest(self, X, X_f, xb=None, xb_f=inf):
		r"""Get the best individual for population.

		Args:
			X (numpy.ndarray): Current population.
			X_f (numpy.ndarray): Current populations fitness/function values of aligned individuals.
			xb (numpy.ndarray): Best individual.
			xb_f (float): Fitness value of best individual.

		Returns:
			Tuple[numpy.ndarray, float]:
				1. Coordinates of best solution.
				2. beset fitness/function value.
		"""
		ib = argmin(X_f)
		if isinstance(X_f, (float, int)) and xb_f >= X_f: xb, xb_f = X, X_f
		elif isinstance(X_f, (ndarray, list)) and xb_f >= X_f[ib]: xb, xb_f = X[ib], X_f[ib]
		return (xb.x.copy() if isinstance(xb, Individual) else xb.copy()), xb_f

	def initPopulation(self, task):
		r"""Initialize starting population of optimization algorithm.

		Args:
			task (Task): Optimization task.

		Returns:
			Tuple[numpy.ndarray, numpy.ndarray, Dict[str, Any]]:
				1. New population.
				2. New population fitness values.
				3. Additional arguments.

		See Also:
			* :func:`NiaPy.algorithms.Algorithm.setParameters`
		"""
		pop, fpop = self.InitPopFunc(task=task, NP=self.NP, rnd=self.Rand, itype=self.itype)
		return pop, fpop, {}

	def runIteration(self, task, pop, fpop, xb, fxb, **dparams):
		r"""Core functionality of algorithm.

		This function is called on every algorithm iteration.

		Args:
			task (Task): Optimization task.
			pop (numpy.ndarray): Current population coordinates.
			fpop (numpy.ndarray): Current population fitness value.
			xb (numpy.ndarray): Current generation best individuals coordinates.
			xb_f (float): current generation best individuals fitness value.
			**dparams (Dict[str, Any]): Additional arguments for algorithms.

		Returns:
			Tuple[numpy.ndarray, numpy.ndarray, numpy.ndarray, float, Dict[str, Any]]:
				1. New populations coordinates.
				2. New populations fitness values.
				3. New global best position/solution
				4. New global best fitness/objective value
				5. Additional arguments of the algorithm.

		See Also:
			* :func:`NiaPy.algorithms.Algorithm.runYield`
		"""
		return pop, fpop, xb, fxb, dparams

	def runYield(self, task):
		r"""Run the algorithm for a single iteration and return the best solution.

		Args:
			task (Task): Task with bounds and objective function for optimization.

		Returns:
			Generator[Tuple[numpy.ndarray, float], None, None]: Generator getting new/old optimal global values.

		Yield:
			Tuple[numpy.ndarray, float]:
				1. New population best individuals coordinates.
				2. Fitness value of the best solution.

		See Also:
			* :func:`NiaPy.algorithms.Algorithm.initPopulation`
			* :func:`NiaPy.algorithms.Algorithm.runIteration`
		"""
		pop, fpop, dparams = self.initPopulation(task)
		xb, fxb = self.getBest(pop, fpop)
		yield xb, fxb
		while True:
			pop, fpop, xb, fxb, dparams = self.runIteration(task, pop, fpop, xb, fxb, **dparams)
			yield xb, fxb

	def runTask(self, task):
		r"""Start the optimization.

		Args:
			task (Task): Task with bounds and objective function for optimization.

		Returns:
			Tuple[numpy.ndarray, float]:
				1. Best individuals components found in optimization process.
				2. Best fitness value found in optimization process.

		See Also:
			* :func:`NiaPy.algorithms.Algorithm.runYield`
		"""
		algo, xb, fxb = self.runYield(task), None, inf
		while not task.stopCond():
			xb, fxb = next(algo)
			task.nextIter()
		return xb, fxb

	def run(self, task):
		r"""Start the optimization.

		Args:
			task (Task): Optimization task.

		Returns:
			Tuple[numpy.ndarray, float]:
				1. Best individuals components found in optimization process.
				2. Best fitness value found in optimization process.

		See Also:
			* :func:`NiaPy.algorithms.Algorithm.runTask`
		"""
		try:
			# task.start()
			r = self.runTask(task)
			return r[0], r[1] * task.optType.value
		except (FesException, GenException, TimeException, RefException): return task.x, task.x_f * task.optType.value
		except Exception as e: self.exception = e
		return None, None

	def bad_run(self):
		r"""Check if some exeptions where thrown when the algorithm was running.

		Returns:
			bool: True if some error where detected at runtime of the algorithm, otherwise False
		"""
		return self.exception is not None

class Individual:
	r"""Class that represents one solution in population of solutions.

	Date:
		2018

	Author:
		Klemen Berkovič

	License:
		MIT

	Attributes:
		x (numpy.ndarray): Coordinates of individual.
		f (float): Function/fitness value of individual.
	"""
	x = None
	f = inf

	def __init__(self, x=None, task=None, e=True, rnd=rand, **kwargs):
		r"""Initialize new individual.

		Parameters:
			task (Optional[Task]): Optimization task.
			rand (Optional[mtrand.RandomState]): Random generator.
			x (Optional[numpy.ndarray]): Individuals components.
			e (Optional[bool]): True to evaluate the individual on initialization. Default value is True.
			**kwargs (Dict[str, Any]): Additional arguments.
		"""
		self.f = task.optType.value * inf if task is not None else inf
		if x is not None: self.x = x if isinstance(x, ndarray) else asarray(x)
		else: self.generateSolution(task, rnd)
		if e and task is not None: self.evaluate(task, rnd)

	def generateSolution(self, task, rnd=rand):
		r"""Generate new solution.

		Generate new solution for this individual and set it to ``self.x``.
		This method uses ``rnd`` for getting random numbers.
		For generating random components ``rnd`` and ``task`` is used.

		Args:
			task (Task): Optimization task.
			rnd (Optional[mtrand.RandomState]): Random numbers generator object.
		"""
		if task is not None: self.x = task.Lower + task.bRange * rnd.rand(task.D)

	def evaluate(self, task, rnd=rand):
		r"""Evaluate the solution.

		Evaluate solution ``this.x`` with the help of task.
		Task is used for reparing the solution and then evaluating it.

		Args:
			task (Task): Objective function object.
			rnd (Optional[mtrand.RandomState]): Random generator.

		See Also:
			* :func:`NiaPy.util.Task.repair`
		"""
		self.x = task.repair(self.x, rnd=rnd)
		self.f = task.eval(self.x)

	def copy(self):
		r"""Return a copy of self.

		Method returns copy of ``this`` object so it is safe for editing.

		Returns:
			Individual: Copy of self.
		"""
		return Individual(x=self.x.copy(), f=self.f, e=False)

	def __eq__(self, other):
		r"""Compare the individuals for equalities.

		Args:
			other (Union[Any, numpy.ndarray]): Object that we want to compare this object to.

		Returns:
			bool: `True` if equal or `False` if no equal.
		"""
		if isinstance(other, ndarray):
			for e in other:
				if self == e: return True
			return False
		return array_equal(self.x, other.x) and self.f == other.f

	def __str__(self):
		r"""Print the individual with the solution and objective value.

		Returns:
			str: String representation of self.
		"""
		return '%s -> %s' % (self.x, self.f)

	def __getitem__(self, i):
		r"""Get the value of i-th component of the solution.

		Args:
			i (int): Position of the solution component.

		Returns:
			Any: Value of ith component.
		"""
		return self.x[i]

	def __setitem__(self, i, v):
		r"""Set the value of i-th component of the solution to v value.

		Args:
			i (int): Position of the solution component.
			v (Any): Value to set to i-th component.
		"""
		self.x[i] = v

	def __len__(self):
		r"""Get the length of the solution or the number of components.

		Returns:
			int: Number of components.
		"""
		return len(self.x)

# vim: tabstop=3 noexpandtab shiftwidth=3 softtabstop=3


1			# encoding=utf8
2			import logging
3
4			from numpy import random as rand, inf, ndarray, asarray, array_equal, argmin, apply_along_axis
5
6			from NiaPy.util import FesException, GenException, TimeException, RefException
7			from NiaPy.util.utility import objects2array
8
9			logging.basicConfig()
10			logger = logging.getLogger('NiaPy.util.utility')
11			logger.setLevel('INFO')
12
13			__all__ = [
14			'Algorithm',
15			'Individual',
16			'defaultIndividualInit',
17			'defaultNumPyInit'
18			]
19
20			def defaultNumPyInit(task, NP, rnd=rand, **kwargs):
21			r"""Initialize starting population that is represented with `numpy.ndarray` with shape `{NP, task.D}`.
22
23			Args:
24			task (Task): Optimization task.
25			NP (int): Number of individuals in population.
26			rnd (Optional[mtrand.RandomState]): Random number generator.
27			kwargs (Dict[str, Any]): Additional arguments.
28
29			Returns:
30			Tuple[numpy.ndarray, numpy.ndarray[float]]:
31			1. New population with shape `{NP, task.D}`.
32			2. New population function/fitness values.
33			"""
34			pop = task.Lower + rnd.rand(NP, task.D) * task.bRange
35			fpop = apply_along_axis(task.eval, 1, pop)
36			return pop, fpop
37
38			def defaultIndividualInit(task, NP, rnd=rand, itype=None, **kwargs):
39			r"""Initialize `NP` individuals of type `itype`.
40
41			Args:
42			task (Task): Optimization task.
43			NP (int): Number of individuals in population.
44			rnd (Optional[mtrand.RandomState]): Random number generator.
45			itype (Optional[Individual]): Class of individual in population.
46			kwargs (Dict[str, Any]): Additional arguments.
47
48			Returns:
49			Tuple[numpy.ndarray[Individual], numpy.ndarray[float]:
50			1. Initialized individuals.
51			2. Initialized individuals function/fitness values.
52			"""
53			pop = objects2array([itype(task=task, rnd=rnd, e=True) for _ in range(NP)])
54			return pop, asarray([x.f for x in pop])
55
56			class Algorithm:
57			r"""Class for implementing algorithms.
58
59			Date:
60			2018
61
62			Author
63			Klemen Berkovič
64
65			License:
66			MIT
67
68			Attributes:
69			Name (List[str]): List of names for algorithm.
70			Rand (mtrand.RandomState): Random generator.
71			NP (int): Number of inidividuals in populatin.
72			InitPopFunc (Callable[[int, Task, mtrand.RandomState, Dict[str, Any]], Tuple[numpy.ndarray, numpy.ndarray[float]]]): Idividual initialization function.
73			itype (Individual): Type of individuals used in population, default value is None for Numpy arrays.
74			"""
75			Name = ['Algorithm', 'AAA']
76			Rand = rand.RandomState(None)
			0 ignored issues – show Comprehensibility Best Practice introduced 2019-04-15 10:29 UTC by Report Bug Copy Issue Report The variable `rand` does not seem to be defined. Loading history...
77			NP = 50
78			InitPopFunc = defaultNumPyInit
79			itype = None
80
81			@staticmethod
82			def typeParameters():
83			r"""Return functions for checking values of parameters.
84
85			Return:
86			Dict[str, Callable]:
87			* NP (Callable[[int], bool]): Check if number of individuals is :math:`\in [0, \infty]`.
88			"""
89			return {'NP': lambda x: isinstance(x, int) and x > 0}
90
91			def __init__(self, **kwargs):
92			r"""Initialize algorithm and create name for an algorithm.
93
94			Args:
95			seed (int): Starting seed for random generator.
96
97			See Also:
98			* :func:`NiaPy.algorithms.Algorithm.setParameters`
99			"""
100			self.Rand, self.exception = rand.RandomState(kwargs.pop('seed', None)), None
101			self.setParameters(**kwargs)
102
103			@staticmethod
104			def algorithmInfo():
105			r"""Get algorithm information.
106
107			Returns:
108			str: Bit item.
109			"""
110			return '''Basic algorithm. No implementation!!!'''
111
112			def setParameters(self, NP=50, InitPopFunc=defaultNumPyInit, itype=None, **kwargs):
113			r"""Set the parameters/arguments of the algorithm.
114
115			Args:
116			NP (Optional[int]): Number of individuals in population :math:`\in [1, \infty]`.
117			InitPopFunc (Optional[Callable[[int, Task, mtrand.RandomState, Dict[str, Any]], Tuple[numpy.ndarray, numpy.ndarray[float]]]]): Type of individuals used by algorithm.
118			itype (Optional[Any]): Individual type used in population, default is Numpy array.
119			**kwargs (Dict[str, Any]): Additional arguments.
120
121			See Also:
122			* :func:`NiaPy.algorithms.defaultNumPyInit`
123			* :func:`NiaPy.algorithms.defaultIndividualInit`
124			"""
125			self.NP, self.InitPopFunc, self.itype = NP, InitPopFunc, itype
126
127			def getParameters(self):
128			r"""Get parameters of the algorithm.
129
130			Returns:
131			Dict[str, Any]:
132			* Parameter name (str): Represents a parameter name
133			* Value of parameter (Any): Represents the value of the parameter
134			"""
135			return {
136			'NP': self.NP,
137			'InitPopFunc': self.InitPopFunc,
138			'itype': self.itype
139			}
140
141			def rand(self, D=1):
142			r"""Get random distribution of shape D in range from 0 to 1.
143
144			Args:
145			D (numpy.ndarray[int]): Shape of returned random distribution.
146
147			Returns:
148			Union[numpy.ndarray[float], float]: Random number or numbers :math:`\in [0, 1]`.
149			"""
150			if isinstance(D, (ndarray, list)): return self.Rand.rand(*D)
151			elif D > 1: return self.Rand.rand(D)
152			else: return self.Rand.rand()
153
154			def uniform(self, Lower, Upper, D=None):
155			r"""Get uniform random distribution of shape D in range from "Lower" to "Upper".
156
157			Args:
158			Lower (Iterable[float]): Lower bound.
159			Upper (Iterable[float]): Upper bound.
160			D (Union[int, Iterable[int]]): Shape of returned uniform random distribution.
161
162			Returns:
163			Union[numpy.ndarray[float], float]: Array of numbers :math:`\in [\mathit{Lower}, \mathit{Upper}]`.
164			"""
165			return self.Rand.uniform(Lower, Upper, D) if D is not None else self.Rand.uniform(Lower, Upper)
166
167			def normal(self, loc, scale, D=None):
168			r"""Get normal random distribution of shape D with mean "loc" and standard deviation "scale".
169
170			Args:
171			loc (float): Mean of the normal random distribution.
172			scale (float): Standard deviation of the normal random distribution.
173			D (Union[int, Iterable[int]]): Shape of returned normal random distribution.
174
175			Returns:
176			Union[numpy.ndarray[float], float]: Array of numbers.
177			"""
178			return self.Rand.normal(loc, scale, D) if D is not None else self.Rand.normal(loc, scale)
179
180			def randn(self, D=None):
181			r"""Get standard normal distribution of shape D.
182
183			Args:
184			D (Union[int, Iterable[int]]): Shape of returned standard normal distribution.
185
186			Returns:
187			Union[numpy.ndarray[float], float]: Random generated numbers or one random generated number :math:`\in [0, 1]`.
188			"""
189			if D is None: return self.Rand.randn()
190			elif isinstance(D, int): return self.Rand.randn(D)
191			return self.Rand.randn(*D)
192
193			def randint(self, Nmax, D=1, Nmin=0, skip=None):
194			r"""Get discrete uniform (integer) random distribution of D shape in range from "Nmin" to "Nmax".
195
196			Args:
197			Nmin (int): Lower integer bound.
198			Nmax (int): One above upper integer bound.
199			D (Union[int, Iterable[int]]): shape of returned discrete uniform random distribution.
200			skip (Union[int, Iterable[int], numpy.ndarray[int]]): numbers to skip.
201
202			Returns:
203			Union[int, numpy.ndarrayj[int]]: Random generated integer number.
204			"""
205			r = None
206			if isinstance(D, (list, tuple, ndarray)): r = self.Rand.randint(Nmin, Nmax, D)
207			elif D > 1: r = self.Rand.randint(Nmin, Nmax, D)
208			else: r = self.Rand.randint(Nmin, Nmax)
209			return r if skip is None or r not in skip else self.randint(Nmax, D, Nmin, skip)
210
211			def getBest(self, X, X_f, xb=None, xb_f=inf):
212			r"""Get the best individual for population.
213
214			Args:
215			X (numpy.ndarray): Current population.
216			X_f (numpy.ndarray): Current populations fitness/function values of aligned individuals.
217			xb (numpy.ndarray): Best individual.
218			xb_f (float): Fitness value of best individual.
219
220			Returns:
221			Tuple[numpy.ndarray, float]:
222			1. Coordinates of best solution.
223			2. beset fitness/function value.
224			"""
225			ib = argmin(X_f)
226			if isinstance(X_f, (float, int)) and xb_f >= X_f: xb, xb_f = X, X_f
227			elif isinstance(X_f, (ndarray, list)) and xb_f >= X_f[ib]: xb, xb_f = X[ib], X_f[ib]
228			return (xb.x.copy() if isinstance(xb, Individual) else xb.copy()), xb_f
229
230			def initPopulation(self, task):
231			r"""Initialize starting population of optimization algorithm.
232
233			Args:
234			task (Task): Optimization task.
235
236			Returns:
237			Tuple[numpy.ndarray, numpy.ndarray, Dict[str, Any]]:
238			1. New population.
239			2. New population fitness values.
240			3. Additional arguments.
241
242			See Also:
243			* :func:`NiaPy.algorithms.Algorithm.setParameters`
244			"""
245			pop, fpop = self.InitPopFunc(task=task, NP=self.NP, rnd=self.Rand, itype=self.itype)
246			return pop, fpop, {}
247
248			def runIteration(self, task, pop, fpop, xb, fxb, **dparams):
249			r"""Core functionality of algorithm.
250
251			This function is called on every algorithm iteration.
252
253			Args:
254			task (Task): Optimization task.
255			pop (numpy.ndarray): Current population coordinates.
256			fpop (numpy.ndarray): Current population fitness value.
257			xb (numpy.ndarray): Current generation best individuals coordinates.
258			xb_f (float): current generation best individuals fitness value.
259			**dparams (Dict[str, Any]): Additional arguments for algorithms.
260
261			Returns:
262			Tuple[numpy.ndarray, numpy.ndarray, numpy.ndarray, float, Dict[str, Any]]:
263			1. New populations coordinates.
264			2. New populations fitness values.
265			3. New global best position/solution
266			4. New global best fitness/objective value
267			5. Additional arguments of the algorithm.
268
269			See Also:
270			* :func:`NiaPy.algorithms.Algorithm.runYield`
271			"""
272			return pop, fpop, xb, fxb, dparams
273
274			def runYield(self, task):
275			r"""Run the algorithm for a single iteration and return the best solution.
276
277			Args:
278			task (Task): Task with bounds and objective function for optimization.
279
280			Returns:
281			Generator[Tuple[numpy.ndarray, float], None, None]: Generator getting new/old optimal global values.
282
283			Yield:
284			Tuple[numpy.ndarray, float]:
285			1. New population best individuals coordinates.
286			2. Fitness value of the best solution.
287
288			See Also:
289			* :func:`NiaPy.algorithms.Algorithm.initPopulation`
290			* :func:`NiaPy.algorithms.Algorithm.runIteration`
291			"""
292			pop, fpop, dparams = self.initPopulation(task)
293			xb, fxb = self.getBest(pop, fpop)
294			yield xb, fxb
295			while True:
296			pop, fpop, xb, fxb, dparams = self.runIteration(task, pop, fpop, xb, fxb, **dparams)
297			yield xb, fxb
298
299			def runTask(self, task):
300			r"""Start the optimization.
301
302			Args:
303			task (Task): Task with bounds and objective function for optimization.
304
305			Returns:
306			Tuple[numpy.ndarray, float]:
307			1. Best individuals components found in optimization process.
308			2. Best fitness value found in optimization process.
309
310			See Also:
311			* :func:`NiaPy.algorithms.Algorithm.runYield`
312			"""
313			algo, xb, fxb = self.runYield(task), None, inf
314			while not task.stopCond():
315			xb, fxb = next(algo)
316			task.nextIter()
317			return xb, fxb
318
319			def run(self, task):
320			r"""Start the optimization.
321
322			Args:
323			task (Task): Optimization task.
324
325			Returns:
326			Tuple[numpy.ndarray, float]:
327			1. Best individuals components found in optimization process.
328			2. Best fitness value found in optimization process.
329
330			See Also:
331			* :func:`NiaPy.algorithms.Algorithm.runTask`
332			"""
333			try:
334			# task.start()
335			r = self.runTask(task)
336			return r[0], r[1] * task.optType.value
337			except (FesException, GenException, TimeException, RefException): return task.x, task.x_f * task.optType.value
338			except Exception as e: self.exception = e
339			return None, None
340
341			def bad_run(self):
342			r"""Check if some exeptions where thrown when the algorithm was running.
343
344			Returns:
345			bool: True if some error where detected at runtime of the algorithm, otherwise False
346			"""
347			return self.exception is not None
348
349			class Individual:
350			r"""Class that represents one solution in population of solutions.
351
352			Date:
353			2018
354
355			Author:
356			Klemen Berkovič
357
358			License:
359			MIT
360
361			Attributes:
362			x (numpy.ndarray): Coordinates of individual.
363			f (float): Function/fitness value of individual.
364			"""
365			x = None
366			f = inf
367
368			def __init__(self, x=None, task=None, e=True, rnd=rand, **kwargs):
369			r"""Initialize new individual.
370
371			Parameters:
372			task (Optional[Task]): Optimization task.
373			rand (Optional[mtrand.RandomState]): Random generator.
374			x (Optional[numpy.ndarray]): Individuals components.
375			e (Optional[bool]): True to evaluate the individual on initialization. Default value is True.
376			**kwargs (Dict[str, Any]): Additional arguments.
377			"""
378			self.f = task.optType.value * inf if task is not None else inf
379			if x is not None: self.x = x if isinstance(x, ndarray) else asarray(x)
380			else: self.generateSolution(task, rnd)
381			if e and task is not None: self.evaluate(task, rnd)
382
383			def generateSolution(self, task, rnd=rand):
384			r"""Generate new solution.
385
386			Generate new solution for this individual and set it to ``self.x``.
387			This method uses ``rnd`` for getting random numbers.
388			For generating random components ``rnd`` and ``task`` is used.
389
390			Args:
391			task (Task): Optimization task.
392			rnd (Optional[mtrand.RandomState]): Random numbers generator object.
393			"""
394			if task is not None: self.x = task.Lower + task.bRange * rnd.rand(task.D)
395
396			def evaluate(self, task, rnd=rand):
397			r"""Evaluate the solution.
398
399			Evaluate solution ``this.x`` with the help of task.
400			Task is used for reparing the solution and then evaluating it.
401
402			Args:
403			task (Task): Objective function object.
404			rnd (Optional[mtrand.RandomState]): Random generator.
405
406			See Also:
407			* :func:`NiaPy.util.Task.repair`
408			"""
409			self.x = task.repair(self.x, rnd=rnd)
410			self.f = task.eval(self.x)
411
412			def copy(self):
413			r"""Return a copy of self.
414
415			Method returns copy of ``this`` object so it is safe for editing.
416
417			Returns:
418			Individual: Copy of self.
419			"""
420			return Individual(x=self.x.copy(), f=self.f, e=False)
421
422			def __eq__(self, other):
423			r"""Compare the individuals for equalities.
424
425			Args:
426			other (Union[Any, numpy.ndarray]): Object that we want to compare this object to.
427
428			Returns:
429			bool: `True` if equal or `False` if no equal.
430			"""
431			if isinstance(other, ndarray):
432			for e in other:
433			if self == e: return True
434			return False
435			return array_equal(self.x, other.x) and self.f == other.f
436
437			def __str__(self):
438			r"""Print the individual with the solution and objective value.
439
440			Returns:
441			str: String representation of self.
442			"""
443			return '%s -> %s' % (self.x, self.f)
444
445			def __getitem__(self, i):
446			r"""Get the value of i-th component of the solution.
447
448			Args:
449			i (int): Position of the solution component.
450
451			Returns:
452			Any: Value of ith component.
453			"""
454			return self.x[i]
455
456			def __setitem__(self, i, v):
457			r"""Set the value of i-th component of the solution to v value.
458
459			Args:
460			i (int): Position of the solution component.
461			v (Any): Value to set to i-th component.
462			"""
463			self.x[i] = v
464
465			def __len__(self):
466			r"""Get the length of the solution or the number of components.
467
468			Returns:
469			int: Number of components.
470			"""
471			return len(self.x)
472
473			# vim: tabstop=3 noexpandtab shiftwidth=3 softtabstop=3
474

NiaOrg / NiaPy

Push — master ( cc7745...279fb5 )

NiaPy.algorithms.algorithm.Individual.evaluate() A

Complexity

Size

Duplication

Importance

Duplication Side-by-Side

Filter issues like