#!/usr/bin/env python
# Created by "Thieu" at 22:07, 11/04/2020 ----------%
# Email: nguyenthieu2102@gmail.com %
# Github: https://github.com/thieu1995 %
# --------------------------------------------------%
import numpy as np
from copy import deepcopy
from mealpy.optimizer import Optimizer
[docs]class BaseVCS(Optimizer):
"""
My changed version of: Virus Colony Search (VCS)
Links:
1. https://doi.org/10.1016/j.advengsoft.2015.11.004
Notes
~~~~~
+ Removes all third loop, makes algrithm 10 times faster than original
+ In Immune response process, updates the whole position instead of updating each variable in position
+ Drops batch-size idea to 3 main process of this algorithm, makes it more robust
Hyper-parameters should fine tuned in approximate range to get faster convergen toward the global optimum:
+ lamda (float): [0.2, 0.5], Number of the best will keep
+ xichma (float): [0.1, 0.5], Weight factor
Examples
~~~~~~~~
>>> import numpy as np
>>> from mealpy.bio_based.VCS import BaseVCS
>>>
>>> def fitness_function(solution):
>>> return np.sum(solution**2)
>>>
>>> problem_dict1 = {
>>> "fit_func": fitness_function,
>>> "lb": [-10, -15, -4, -2, -8],
>>> "ub": [10, 15, 12, 8, 20],
>>> "minmax": "min",
>>> "verbose": True,
>>> }
>>>
>>> epoch = 1000
>>> pop_size = 50
>>> lamda = 0.5
>>> xichma = 0.3
>>> model = BaseVCS(problem_dict1, epoch, pop_size, lamda, xichma)
>>> best_position, best_fitness = model.solve()
>>> print(f"Solution: {best_position}, Fitness: {best_fitness}")
"""
def __init__(self, problem, epoch=10000, pop_size=100, lamda=0.5, xichma=0.3, **kwargs):
"""
Args:
problem (dict): The problem dictionary
epoch (int): maximum number of iterations, default = 10000
pop_size (int): number of population size, default = 100
lamda (float): Number of the best will keep, default = 0.5
xichma (float): Weight factor, default = 0.3
"""
super().__init__(problem, kwargs)
self.nfe_per_epoch = 3 * pop_size
self.sort_flag = True
self.epoch = epoch
self.pop_size = pop_size
self.xichma = xichma
self.lamda = lamda
if lamda < 1:
self.n_best = int(lamda * self.pop_size)
else:
self.n_best = int(lamda)
def _calculate_xmean(self, pop):
"""
Calculate the mean position of list of solutions (population)
Args:
pop (list): List of solutions (population)
Returns:
list: Mean position
"""
## Calculate the weighted mean of the λ best individuals by
pop, local_best = self.get_global_best_solution(pop)
pos_list = [agent[self.ID_POS] for agent in pop[:self.n_best]]
factor_down = self.n_best * np.log1p(self.n_best + 1) - np.log1p(np.prod(range(1, self.n_best + 1)))
weight = np.log1p(self.n_best + 1) / factor_down
weight = weight / self.n_best
x_mean = weight * np.sum(pos_list, axis=0)
return x_mean
[docs] def evolve(self, epoch):
"""
The main operations (equations) of algorithm. Inherit from Optimizer class
Args:
epoch (int): The current iteration
"""
## Viruses diffusion
for i in range(0, self.pop_size):
xichma = (np.log1p(epoch + 1) / self.epoch) * (self.pop[i][self.ID_POS] - self.g_best[self.ID_POS])
gauss = np.random.normal(np.random.normal(self.g_best[self.ID_POS], np.abs(xichma)))
pos_new = gauss + np.random.uniform() * self.g_best[self.ID_POS] - np.random.uniform() * self.pop[i][self.ID_POS]
self.pop[i][self.ID_POS] = self.amend_position(pos_new)
self.pop = self.update_fitness_population(self.pop)
## Host cells infection
x_mean = self._calculate_xmean(self.pop)
xichma = self.xichma * (1 - (epoch + 1) / self.epoch)
for i in range(0, self.pop_size):
## Basic / simple version, not the original version in the paper
pos_new = x_mean + xichma * np.random.normal(0, 1, self.problem.n_dims)
self.pop[i][self.ID_POS] = self.amend_position(pos_new)
self.pop = self.update_fitness_population(self.pop)
## Calculate the weighted mean of the λ best individuals by
self.pop, g_best = self.get_global_best_solution(self.pop)
## Immune response
for i in range(0, self.pop_size):
pr = (self.problem.n_dims - i + 1) / self.problem.n_dims
id1, id2 = np.random.choice(list(set(range(0, self.pop_size)) - {i}), 2, replace=False)
temp = self.pop[id1][self.ID_POS] - (self.pop[id2][self.ID_POS] - self.pop[i][self.ID_POS]) * np.random.uniform()
pos_new = np.where(np.random.uniform(0, 1, self.problem.n_dims) < pr, self.pop[i][self.ID_POS], temp)
self.pop[i][self.ID_POS] = self.amend_position(pos_new)
self.pop = self.update_fitness_population(self.pop)
[docs]class OriginalVCS(BaseVCS):
"""
The original version of: Virus Colony Search (VCS)
Links:
1. https://doi.org/10.1016/j.advengsoft.2015.11.004
Notes
~~~~~
This is basic version, not the full version of the paper
Hyper-parameters should fine tuned in approximate range to get faster convergen toward the global optimum:
+ lamda (float): [0.2, 0.5], Number of the best will keep
+ xichma (float): [0.1, 0.5], Weight factor
Examples
~~~~~~~~
>>> import numpy as np
>>> from mealpy.bio_based.VCS import OriginalVCS
>>>
>>> def fitness_function(solution):
>>> return np.sum(solution**2)
>>>
>>> problem_dict1 = {
>>> "fit_func": fitness_function,
>>> "lb": [-10, -15, -4, -2, -8],
>>> "ub": [10, 15, 12, 8, 20],
>>> "minmax": "min",
>>> "verbose": True,
>>> }
>>>
>>> epoch = 1000
>>> pop_size = 50
>>> lamda = 0.5
>>> xichma = 0.3
>>> model = OriginalVCS(problem_dict1, epoch, pop_size, lamda, xichma)
>>> best_position, best_fitness = model.solve()
>>> print(f"Solution: {best_position}, Fitness: {best_fitness}")
References
~~~~~~~~~~
[1] Li, M.D., Zhao, H., Weng, X.W. and Han, T., 2016. A novel nature-inspired algorithm
for optimization: Virus colony search. Advances in Engineering Software, 92, pp.65-88.
"""
def __init__(self, problem, epoch=10000, pop_size=100, lamda=0.5, xichma=0.3, **kwargs):
"""
Args:
problem (dict): The problem dictionary
epoch (int): maximum number of iterations, default = 10000
pop_size (int): number of population size, default = 100
lamda (float): Number of the best will keep, default = 0.5
xichma (float): Weight factor, default = 0.3
"""
super().__init__(problem, epoch, pop_size, lamda, xichma, **kwargs)
[docs] def amend_position(self, position):
"""
If solution out of bound at dimension x, then it will re-arrange to random location in the range of domain
Args:
position: vector position (location) of the solution.
Returns:
Amended position
"""
return np.where(np.logical_and(self.problem.lb <= position, position <= self.problem.ub),
position, np.random.uniform(self.problem.lb, self.problem.ub))
[docs] def evolve(self, epoch):
"""
The main operations (equations) of algorithm. Inherit from Optimizer class
Args:
epoch (int): The current iteration
"""
pop = deepcopy(self.pop)
## Viruses diffusion
for i in range(0, self.pop_size):
xichma = (np.log1p(epoch + 1) / self.epoch) * (pop[i][self.ID_POS] - self.g_best[self.ID_POS])
gauss = np.array([np.random.normal(self.g_best[self.ID_POS][idx], np.abs(xichma[idx])) for idx in range(0, self.problem.n_dims)])
pos_new = gauss + np.random.uniform() * self.g_best[self.ID_POS] - np.random.uniform() * pop[i][self.ID_POS]
pop[i][self.ID_POS] = self.amend_position(pos_new)
pop = self.update_fitness_population(pop)
## Host cells infection
x_mean = self._calculate_xmean(pop)
xichma = self.xichma * (1 - (epoch + 1) / self.epoch)
for i in range(0, self.pop_size):
## Basic / simple version, not the original version in the paper
pos_new = x_mean + xichma * np.random.normal(0, 1, self.problem.n_dims)
pop[i][self.ID_POS] = self.amend_position(pos_new)
pop = self.update_fitness_population(pop)
## Immune response
for i in range(0, self.pop_size):
pr = (self.problem.n_dims - i + 1) / self.problem.n_dims
pos_new = pop[i][self.ID_POS]
for j in range(0, self.problem.n_dims):
if np.random.uniform() > pr:
id1, id2 = np.random.choice(list(set(range(0, self.pop_size)) - {i}), 2, replace=False)
pos_new[j] = pop[id1][self.ID_POS][j] - (pop[id2][self.ID_POS][j] - pop[i][self.ID_POS][j]) * np.random.uniform()
pop[i][self.ID_POS] = self.amend_position(pos_new)
pop = self.update_fitness_population(pop)
## Greedy selection
for idx in range(0, self.pop_size):
if self.compare_agent(pop[idx], self.pop[idx]):
self.pop[idx] = deepcopy(pop[idx])