NiaPy.algorithms.basic.hho.HarrisHawksOptimization.initPopulation() - Code Metrics - Inspection of "Harris Hawks Optimization integration" - NiaOrg/NiaPy - Measure and Improve Code Quality continuously with Scrutinizer

Passed

Pull Request — master (#280)

unknown

created 2020-11-15 17:39 UTC

HarrisHawksOptimization.initPopulation() A

↳ Parent: NiaPy.algorithms.basic.hho

Complexity

Conditions

Size

Total Lines	16
Code Lines	3

Duplication

Lines	0
Ratio	0 %

Importance

Changes

Metric	Value
cc	1
eloc	3
nop	2
dl	0
loc	16
rs	10
c	0
b	0
f	0

# encoding=utf8
import logging

from numpy import random as rand, sin, pi, argmin, abs, mean
from scipy.special import gamma

from NiaPy.algorithms.algorithm import Algorithm

logging.basicConfig()
logger = logging.getLogger('NiaPy.algorithms.basic')
logger.setLevel('INFO')

__all__ = ['HarrisHawksOptimization']


def levy_function(dims, step=0.01, rnd=rand):
    r"""Calculate levy function.

    Parameters:
            dim (int): Number of dimensions
            step (float): Step of the Levy function

    Returns:
            float: The Levy function evaluation
    """
    beta = 1.5
    sigma = (gamma(1 + beta) * sin(pi * beta / 2) / (gamma((1 + beta / 2) * beta * 2.0 ** ((beta - 1) / 2)))) ** (1 / beta)
    normal_1 = rnd.normal(0, sigma, size=dims)
    normal_2 = rnd.normal(0, 1, size=dims)
    result = step * normal_1 / (abs(normal_2) ** (1 / beta))
    return result


class HarrisHawksOptimization(Algorithm):
    r"""Implementation of Harris Hawks Optimization algorithm.

    Algorithm:
            Harris Hawks Optimization

    Date:
            2020

    Authors:
            Francisco Jose Solis-Munoz

    License:
            MIT

    Reference paper:
            Heidari et al. "Harris hawks optimization: Algorithm and applications". Future Generation Computer Systems. 2019. Vol. 97. 849-872.

    Attributes:
            Name (List[str]): List of strings representing algorithm name.
            levy (float): Levy factor.

    See Also:
            * :class:`NiaPy.algorithms.Algorithm`
    """
    Name = ['HarrisHawksOptimization', 'HHO']

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

        Returns:
                str: Algorithm information.

        See Also:
                * :func:`NiaPy.algorithms.Algorithm.algorithmInfo`
        """
        return r"""Heidari et al. "Harris hawks optimization: Algorithm and applications". Future Generation Computer Systems. 2019. Vol. 97. 849-872."""

    @staticmethod
    def typeParameters():
        r"""Return dict with where key of dict represents parameter name and values represent checking functions for selected parameter.

        Returns:
                Dict[str, Callable]:
                        * levy (Callable[[Union[float, int]], bool]): Levy factor.

        See Also:
                * :func:`NiaPy.algorithms.Algorithm.typeParameters`
        """
        d = Algorithm.typeParameters()
        d.update({
            'levy': lambda x: isinstance(x, (float, int)) and x > 0,
        })
        return d

    def setParameters(self, NP=40, levy=0.01, **ukwargs):
        r"""Set the parameters of the algorithm.

        Args:
                levy (Optional[float]): Levy factor.

        See Also:
                * :func:`NiaPy.algorithms.Algorithm.setParameters`
        """
        Algorithm.setParameters(self, NP=NP, **ukwargs)
        self.levy = levy

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

        Returns:
                Dict[str, Any]
        """
        d = Algorithm.getParameters(self)
        d.update({
            'levy': self.levy
        })
        return d

    def initPopulation(self, task):
        r"""Initialize the starting population.

        Parameters:
                task (Task): Optimization task

        Returns:
                Tuple[numpy.ndarray, numpy.ndarray[float], Dict[str, Any]]:
                        1. New population.
                        2. New population fitness/function values.

        See Also:
                * :func:`NiaPy.algorithms.Algorithm.initPopulation`
        """
        Sol, Fitness, d = Algorithm.initPopulation(self, task)
        return Sol, Fitness, d

    def runIteration(self, task, Sol, Fitness, xb, fxb, **dparams):
        r"""Core function of Harris Hawks Optimization.

        Parameters:
                task (Task): Optimization task.
                Sol (numpy.ndarray): Current population
                Fitness (numpy.ndarray[float]): Current population fitness/funciton values
                xb (numpy.ndarray): Current best individual
                fxb (float): Current best individual function/fitness value
                dparams (Dict[str, Any]): Additional algorithm arguments

        Returns:
                Tuple[numpy.ndarray, numpy.ndarray, numpy.ndarray, float, Dict[str, Any]]:
                        1. New population
                        2. New population fitness/function vlues
                        3. New global best solution
                        4. New global best fitness/objective value
        """
        rnd = self.Rand
        # Decreasing energy factor
        decreasing_energy_factor = 2 * (1 - task.iters() / task.nGEN)
        mean_sol = mean(Sol)
        # Update population
        for i in range(self.NP):
            jumping_energy = rnd.uniform(0, 2)
            decreasing_energy_random = rnd.uniform(-1, 1)
            escaping_energy = decreasing_energy_factor * decreasing_energy_random
            escaping_energy_abs = abs(escaping_energy)
            random_number = rnd.rand()
            if escaping_energy >= 1 and random_number >= 0.5:
                # 0. Exploration: Random tall tree
                rhi = rnd.randint(0, self.NP)
                random_agent = Sol[rhi]
                Sol[i] = random_agent - rnd.rand() * abs(random_agent - 2 * rnd.rand() * Sol[i])
            elif escaping_energy_abs >= 1 and random_number < 0.5:
                # 1. Exploration: Family members mean
                Sol[i] = (xb - mean_sol) - rnd.rand() * rnd.uniform(task.Lower, task.Upper)
            elif escaping_energy_abs >= 0.5 and random_number >= 0.5:
                # 2. Exploitation: Soft besiege
                Sol[i] = \
                    (xb - Sol[i]) - \
                    escaping_energy * \
                    abs(jumping_energy * xb - Sol[i])
            elif escaping_energy_abs < 0.5 and random_number >= 0.5:
                # 3. Exploitation: Hard besiege
                Sol[i] = \
                    xb - \
                    escaping_energy * \
                    abs(xb - Sol[i])
            elif escaping_energy_abs >= 0.5 and random_number < 0.5:
                # 4. Exploitation: Soft besiege with pprogressive rapid dives
                cand1 = task.repair(xb - escaping_energy * abs(jumping_energy * xb - Sol[i]), rnd=rand)
                random_vector = rnd.rand(task.D)
                cand2 = task.repair(cand1 + random_vector * levy_function(task.D, self.levy, rnd=rand), rnd=rand)
                if task.eval(cand1) < Fitness[i]:
                    Sol[i] = cand1
                elif task.eval(cand2) < Fitness[i]:
                    Sol[i] = cand2
            elif escaping_energy_abs < 0.5 and random_number < 0.5:
                # 5. Exploitation: Hard besiege with progressive rapid dives
                cand1 = task.repair(xb - escaping_energy * abs(jumping_energy * xb - mean_sol), rnd=rand)
                random_vector = rnd.rand(task.D)
                cand2 = task.repair(cand1 + random_vector * levy_function(task.D, self.levy, rnd=rand), rnd=rand)
                if task.eval(cand1) < Fitness[i]:
                    Sol[i] = cand1
                elif task.eval(cand2) < Fitness[i]:
                    Sol[i] = cand2
            # Repair agent (from population) values
            Sol[i] = task.repair(Sol[i], rnd=rand)
            # Eval population
            Fitness[i] = task.eval(Sol[i])
        # Get best of population
        best_index = argmin(Fitness)
        xb_cand = Sol[best_index].copy()
        fxb_cand = Fitness[best_index].copy()
        if fxb_cand < fxb:
            fxb = fxb_cand
            xb = xb_cand.copy()
        return Sol, Fitness, xb, fxb, {}

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


1			# encoding=utf8
2			import logging
3
4			from numpy import random as rand, sin, pi, argmin, abs, mean
5			from scipy.special import gamma
6
7			from NiaPy.algorithms.algorithm import Algorithm
8
9			logging.basicConfig()
10			logger = logging.getLogger('NiaPy.algorithms.basic')
11			logger.setLevel('INFO')
12
13			__all__ = ['HarrisHawksOptimization']
14
15
16			def levy_function(dims, step=0.01, rnd=rand):
17			r"""Calculate levy function.
18
19			Parameters:
20			dim (int): Number of dimensions
21			step (float): Step of the Levy function
22
23			Returns:
24			float: The Levy function evaluation
25			"""
26			beta = 1.5
27			sigma = (gamma(1 + beta) * sin(pi * beta / 2) / (gamma((1 + beta / 2) * beta * 2.0 ((beta - 1) / 2)))) (1 / beta)
28			normal_1 = rnd.normal(0, sigma, size=dims)
29			normal_2 = rnd.normal(0, 1, size=dims)
30			result = step * normal_1 / (abs(normal_2) ** (1 / beta))
31			return result
32
33
34			class HarrisHawksOptimization(Algorithm):
35			r"""Implementation of Harris Hawks Optimization algorithm.
36
37			Algorithm:
38			Harris Hawks Optimization
39
40			Date:
41			2020
42
43			Authors:
44			Francisco Jose Solis-Munoz
45
46			License:
47			MIT
48
49			Reference paper:
50			Heidari et al. "Harris hawks optimization: Algorithm and applications". Future Generation Computer Systems. 2019. Vol. 97. 849-872.
51
52			Attributes:
53			Name (List[str]): List of strings representing algorithm name.
54			levy (float): Levy factor.
55
56			See Also:
57			* :class:`NiaPy.algorithms.Algorithm`
58			"""
59			Name = ['HarrisHawksOptimization', 'HHO']
60
61			@staticmethod
62			def algorithmInfo():
63			r"""Get algorithms information.
64
65			Returns:
66			str: Algorithm information.
67
68			See Also:
69			* :func:`NiaPy.algorithms.Algorithm.algorithmInfo`
70			"""
71			return r"""Heidari et al. "Harris hawks optimization: Algorithm and applications". Future Generation Computer Systems. 2019. Vol. 97. 849-872."""
72
73			@staticmethod
74			def typeParameters():
75			r"""Return dict with where key of dict represents parameter name and values represent checking functions for selected parameter.
76
77			Returns:
78			Dict[str, Callable]:
79			* levy (Callable[[Union[float, int]], bool]): Levy factor.
80
81			See Also:
82			* :func:`NiaPy.algorithms.Algorithm.typeParameters`
83			"""
84			d = Algorithm.typeParameters()
85			d.update({
86			'levy': lambda x: isinstance(x, (float, int)) and x > 0,
87			})
88			return d
89
90			def setParameters(self, NP=40, levy=0.01, **ukwargs):
91			r"""Set the parameters of the algorithm.
92
93			Args:
94			levy (Optional[float]): Levy factor.
95
96			See Also:
97			* :func:`NiaPy.algorithms.Algorithm.setParameters`
98			"""
99			Algorithm.setParameters(self, NP=NP, **ukwargs)
100			self.levy = levy
101
102			def getParameters(self):
103			r"""Get parameters of the algorithm.
104
105			Returns:
106			Dict[str, Any]
107			"""
108			d = Algorithm.getParameters(self)
109			d.update({
110			'levy': self.levy
111			})
112			return d
113
114			def initPopulation(self, task):
115			r"""Initialize the starting population.
116
117			Parameters:
118			task (Task): Optimization task
119
120			Returns:
121			Tuple[numpy.ndarray, numpy.ndarray[float], Dict[str, Any]]:
122			1. New population.
123			2. New population fitness/function values.
124
125			See Also:
126			* :func:`NiaPy.algorithms.Algorithm.initPopulation`
127			"""
128			Sol, Fitness, d = Algorithm.initPopulation(self, task)
129			return Sol, Fitness, d
130
131			def runIteration(self, task, Sol, Fitness, xb, fxb, **dparams):
132			r"""Core function of Harris Hawks Optimization.
133
134			Parameters:
135			task (Task): Optimization task.
136			Sol (numpy.ndarray): Current population
137			Fitness (numpy.ndarray[float]): Current population fitness/funciton values
138			xb (numpy.ndarray): Current best individual
139			fxb (float): Current best individual function/fitness value
140			dparams (Dict[str, Any]): Additional algorithm arguments
141
142			Returns:
143			Tuple[numpy.ndarray, numpy.ndarray, numpy.ndarray, float, Dict[str, Any]]:
144			1. New population
145			2. New population fitness/function vlues
146			3. New global best solution
147			4. New global best fitness/objective value
148			"""
149			rnd = self.Rand
150			# Decreasing energy factor
151			decreasing_energy_factor = 2 * (1 - task.iters() / task.nGEN)
152			mean_sol = mean(Sol)
153			# Update population
154			for i in range(self.NP):
155			jumping_energy = rnd.uniform(0, 2)
156			decreasing_energy_random = rnd.uniform(-1, 1)
157			escaping_energy = decreasing_energy_factor * decreasing_energy_random
158			escaping_energy_abs = abs(escaping_energy)
159			random_number = rnd.rand()
160			if escaping_energy >= 1 and random_number >= 0.5:
161			# 0. Exploration: Random tall tree
162			rhi = rnd.randint(0, self.NP)
163			random_agent = Sol[rhi]
164			Sol[i] = random_agent - rnd.rand() * abs(random_agent - 2 * rnd.rand() * Sol[i])
165			elif escaping_energy_abs >= 1 and random_number < 0.5:
166			# 1. Exploration: Family members mean
167			Sol[i] = (xb - mean_sol) - rnd.rand() * rnd.uniform(task.Lower, task.Upper)
168			elif escaping_energy_abs >= 0.5 and random_number >= 0.5:
169			# 2. Exploitation: Soft besiege
170			Sol[i] = \
171			(xb - Sol[i]) - \
172			escaping_energy * \
173			abs(jumping_energy * xb - Sol[i])
174			elif escaping_energy_abs < 0.5 and random_number >= 0.5:
175			# 3. Exploitation: Hard besiege
176			Sol[i] = \
177			xb - \
178			escaping_energy * \
179			abs(xb - Sol[i])
180			elif escaping_energy_abs >= 0.5 and random_number < 0.5:
181			# 4. Exploitation: Soft besiege with pprogressive rapid dives
182			cand1 = task.repair(xb - escaping_energy * abs(jumping_energy * xb - Sol[i]), rnd=rand)
183			random_vector = rnd.rand(task.D)
184			cand2 = task.repair(cand1 + random_vector * levy_function(task.D, self.levy, rnd=rand), rnd=rand)
185			if task.eval(cand1) < Fitness[i]:
186			Sol[i] = cand1
187			elif task.eval(cand2) < Fitness[i]:
188			Sol[i] = cand2
189			elif escaping_energy_abs < 0.5 and random_number < 0.5:
190			# 5. Exploitation: Hard besiege with progressive rapid dives
191			cand1 = task.repair(xb - escaping_energy * abs(jumping_energy * xb - mean_sol), rnd=rand)
192			random_vector = rnd.rand(task.D)
193			cand2 = task.repair(cand1 + random_vector * levy_function(task.D, self.levy, rnd=rand), rnd=rand)
194			if task.eval(cand1) < Fitness[i]:
195			Sol[i] = cand1
196			elif task.eval(cand2) < Fitness[i]:
197			Sol[i] = cand2
198			# Repair agent (from population) values
199			Sol[i] = task.repair(Sol[i], rnd=rand)
200			# Eval population
201			Fitness[i] = task.eval(Sol[i])
202			# Get best of population
203			best_index = argmin(Fitness)
204			xb_cand = Sol[best_index].copy()
205			fxb_cand = Fitness[best_index].copy()
206			if fxb_cand < fxb:
207			fxb = fxb_cand
208			xb = xb_cand.copy()
209			return Sol, Fitness, xb, fxb, {}
210
211			# vim: tabstop=3 noexpandtab shiftwidth=3 softtabstop=3
212

NiaOrg / NiaPy

Pull Request — master (#280)

HarrisHawksOptimization.initPopulation() A

Complexity

Size

Duplication

Importance

Duplication Side-by-Side

Filter issues like