#!/usr/bin/env python
# Created by "Thieu" at 07:03, 18/03/2020 ----------%
# Email: nguyenthieu2102@gmail.com %
# Github: https://github.com/thieu1995 %
# --------------------------------------------------%
import numpy as np
from mealpy.optimizer import Optimizer
from mealpy.utils.agent import Agent
[docs]class OriginalASO(Optimizer):
"""
The original version of: Atom Search Optimization (ASO)
Links:
1. https://doi.org/10.1016/j.knosys.2018.08.030
2. https://www.mathworks.com/matlabcentral/fileexchange/67011-atom-search-optimization-aso-algorithm
Hyper-parameters should fine-tune in approximate range to get faster convergence toward the global optimum:
+ alpha (int): Depth weight, default = 10, depend on the problem
+ beta (float): Multiplier weight, default = 0.2
Examples
~~~~~~~~
>>> import numpy as np
>>> from mealpy import FloatVar, ASO
>>>
>>> def objective_function(solution):
>>> return np.sum(solution**2)
>>>
>>> problem_dict = {
>>> "bounds": FloatVar(n_vars=30, lb=(-10.,) * 30, ub=(10.,) * 30, name="delta"),
>>> "minmax": "min",
>>> "obj_func": objective_function
>>> }
>>>
>>> model = ASO.OriginalASO(epoch=1000, pop_size=50, alpha = 50, beta = 0.2)
>>> g_best = model.solve(problem_dict)
>>> print(f"Solution: {g_best.solution}, Fitness: {g_best.target.fitness}")
>>> print(f"Solution: {model.g_best.solution}, Fitness: {model.g_best.target.fitness}")
References
~~~~~~~~~~
[1] Zhao, W., Wang, L. and Zhang, Z., 2019. Atom search optimization and its application to solve a
hydrogeologic parameter estimation problem. Knowledge-Based Systems, 163, pp.283-304.
"""
def __init__(self, epoch: int = 10000, pop_size: int = 100, alpha: int = 10, beta: float = 0.2, **kwargs: object) -> None:
"""
Args:
epoch (int): maximum number of iterations, default = 10000
pop_size (int): number of population size, default = 100
alpha (int): [2, 20], Depth weight, default = 10
beta (float): [0.1, 1.0], Multiplier weight, default = 0.2
"""
super().__init__(**kwargs)
self.epoch = self.validator.check_int("epoch", epoch, [1, 100000])
self.pop_size = self.validator.check_int("pop_size", pop_size, [5, 10000])
self.alpha = self.validator.check_int("alpha", alpha, [1, 100])
self.beta = self.validator.check_float("beta", beta, (0, 1.0))
self.set_parameters(["epoch", "pop_size", "alpha", "beta"])
self.sort_flag = False
[docs] def generate_empty_agent(self, solution: np.ndarray = None) -> Agent:
if solution is None:
solution = self.problem.generate_solution(encoded=True)
velocity = self.generator.uniform(self.problem.lb, self.problem.ub)
mass = 0.0
return Agent(solution=solution, velocity=velocity, mass=mass)
[docs] def amend_solution(self, solution: np.ndarray) -> np.ndarray:
condition = np.logical_and(self.problem.lb <= solution, solution <= self.problem.ub)
rand_pos = self.generator.uniform(self.problem.lb, self.problem.ub)
return np.where(condition, solution, rand_pos)
[docs] def update_mass__(self, population):
list_fit = np.array([agent.target.fitness for agent in population])
list_fit = np.exp(-(list_fit - np.max(list_fit)) / (np.max(list_fit) - np.min(list_fit) + self.EPSILON))
list_fit = list_fit / np.sum(list_fit)
for idx in range(0, self.pop_size):
population[idx].mass = list_fit[idx]
return population
[docs] def find_LJ_potential__(self, iteration, average_dist, radius):
c = (1 - iteration / self.epoch) ** 3
# g0 = 1.1, u = 2.4
rsmin = 1.1 + 0.1 * np.sin(iteration / self.epoch * np.pi / 2)
rsmax = 1.24
if radius / average_dist < rsmin:
rs = rsmin
else:
if radius / average_dist > rsmax:
rs = rsmax
else:
rs = radius / average_dist
potential = c * (12 * (-rs) ** (-13) - 6 * (-rs) ** (-7))
return potential
[docs] def acceleration__(self, population, g_best, iteration):
eps = 2 ** (-52)
pop = self.update_mass__(population)
G = np.exp(-20.0 * iteration / self.epoch)
k_best = int(self.pop_size - (self.pop_size - 2) * (iteration / self.epoch) ** 0.5) + 1
if self.problem.minmax == "min":
k_best_pop = sorted(pop, key=lambda agent: agent.mass, reverse=True)[:k_best].copy()
else:
k_best_pop = sorted(pop, key=lambda agent: agent.mass)[:k_best].copy()
mk_average = np.mean([agent.solution for agent in k_best_pop])
acc_list = np.zeros((self.pop_size, self.problem.n_dims))
for idx in range(0, self.pop_size):
dist_average = np.linalg.norm(pop[idx].solution - mk_average)
temp = np.zeros((self.problem.n_dims))
for atom in k_best_pop:
# calculate LJ-potential
radius = np.linalg.norm(pop[idx].solution - atom.solution)
potential = self.find_LJ_potential__(iteration, dist_average, radius)
temp += potential * self.generator.uniform(0, 1, self.problem.n_dims) * ((atom.solution - pop[idx].solution) / (radius + eps))
temp = self.alpha * temp + self.beta * (g_best.solution - pop[idx].solution)
# calculate acceleration
acc = G * temp / pop[idx].mass
acc_list[idx] = acc
return acc_list
[docs] def evolve(self, epoch):
"""
The main operations (equations) of algorithm. Inherit from Optimizer class
Args:
epoch (int): The current iteration
"""
# Calculate acceleration.
atom_acc_list = self.acceleration__(self.pop, self.g_best, iteration=epoch)
# Update velocity based on random dimensions and position of global best
pop_new = []
for idx in range(0, self.pop_size):
agent = self.pop[idx].copy()
velocity = self.generator.random(self.problem.n_dims) * self.pop[idx].velocity + atom_acc_list[idx]
pos_new = self.pop[idx].solution + velocity
# Relocate atom out of range
pos_new = self.correct_solution(pos_new)
agent.solution = pos_new
pop_new.append(agent)
if self.mode not in self.AVAILABLE_MODES:
agent.target = self.get_target(pos_new)
self.pop[idx] = self.get_better_agent(agent, self.pop[idx], self.problem.minmax)
if self.mode in self.AVAILABLE_MODES:
pop_new = self.update_target_for_population(pop_new)
self.pop = self.greedy_selection_population(self.pop, pop_new, self.problem.minmax)
current_best = self.get_best_agent(pop_new, self.problem.minmax)
if self.compare_target(self.g_best.target, current_best.target, self.problem.minmax):
self.pop[self.generator.integers(0, self.pop_size)] = self.g_best.copy()