Source code for mealpy.physics_based.ASO

# !/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 copy import deepcopy
from mealpy.optimizer import Optimizer


[docs]class BaseASO(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 tuned in approximate range to get faster convergen toward the global optimum: + alpha (int): Depth weight, default = 50 + beta (float): Multiplier weight, default = 0.2 Examples ~~~~~~~~ >>> import numpy as np >>> from mealpy.physics_based.ASO import BaseASO >>> >>> 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 >>> alpha = 50 >>> beta = 0.2 >>> model = BaseASO(problem_dict1, epoch, pop_size, alpha, beta) >>> best_position, best_fitness = model.solve() >>> print(f"Solution: {best_position}, Fitness: {best_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. """ ID_POS = 0 ID_TAR = 1 ID_VEL = 2 # Velocity ID_MAS = 3 # Mass of atom def __init__(self, problem, epoch=10000, pop_size=100, alpha=50, beta=0.2, **kwargs): """ Args: problem (dict): The problem dictionary epoch (int): maximum number of iterations, default = 10000 pop_size (int): number of population size, default = 100 alpha (int): [10, 100], Depth weight, default = 50 beta (float): [0.1, 1.0], Multiplier weight, default = 0.2 """ super().__init__(problem, kwargs) self.nfe_per_epoch = pop_size self.sort_flag = False self.epoch = epoch self.pop_size = pop_size self.alpha = alpha self.beta = beta
[docs] def create_solution(self): """ To get the position, fitness wrapper, target and obj list + A[self.ID_POS] --> Return: position + A[self.ID_TAR] --> Return: [target, [obj1, obj2, ...]] + A[self.ID_TAR][self.ID_FIT] --> Return: target + A[self.ID_TAR][self.ID_OBJ] --> Return: [obj1, obj2, ...] Returns: list: wrapper of solution with format [position, [target, [obj1, obj2, ...]], velocity, mass] """ position = np.random.uniform(self.problem.lb, self.problem.ub) position = self.amend_position(position) fitness = self.get_fitness_position(position=position) velocity = np.random.uniform(self.problem.lb, self.problem.ub) mass = 0.0 return [position, fitness, velocity, mass]
[docs] def amend_position(self, position=None): """ 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))
def _update_mass__(self, population): fit_total, fit_best, fit_worst = self.get_special_fitness(population) for agent in population: agent[self.ID_MAS] = np.exp((agent[self.ID_TAR][self.ID_FIT] - fit_best) / (fit_worst - fit_best + self.EPSILON)) / fit_total return population 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 + 1) / 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 def _acceleration__(self, population, g_best, iteration): eps = 2 ** (-52) pop = self._update_mass__(population) G = np.exp(-20.0 * (iteration + 1) / self.epoch) k_best = int(self.pop_size - (self.pop_size - 2) * ((iteration + 1) / self.epoch) ** 0.5) + 1 if self.problem.minmax == "min": k_best_pop = deepcopy(sorted(pop, key=lambda agent: agent[self.ID_MAS], reverse=True)[:k_best]) else: k_best_pop = deepcopy(sorted(pop, key=lambda agent: agent[self.ID_MAS])[:k_best]) mk_average = np.mean([item[self.ID_POS] for item in k_best_pop]) acc_list = np.zeros((self.pop_size, self.problem.n_dims)) for i in range(0, self.pop_size): dist_average = np.linalg.norm(pop[i][self.ID_POS] - mk_average) temp = np.zeros((self.problem.n_dims)) for atom in k_best_pop: # calculate LJ-potential radius = np.linalg.norm(pop[i][self.ID_POS] - atom[self.ID_POS]) potential = self._find_LJ_potential__(iteration, dist_average, radius) temp += potential * np.random.uniform(0, 1, self.problem.n_dims) * ((atom[self.ID_POS] - pop[i][self.ID_POS]) / (radius + eps)) temp = self.alpha * temp + self.beta * (g_best[self.ID_POS] - pop[i][self.ID_POS]) # calculate acceleration acc = G * temp / pop[i][self.ID_MAS] acc_list[i] = 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 = deepcopy(self.pop[idx]) velocity_rand = np.random.uniform(self.problem.lb, self.problem.ub) velocity = velocity_rand * self.pop[idx][self.ID_VEL] + atom_acc_list[idx] pos_new = self.pop[idx][self.ID_POS] + velocity # Relocate atom out of range agent[self.ID_POS] = self.amend_position(pos_new) pop_new.append(agent) pop_new = self.update_fitness_population(pop_new) pop_new = self.greedy_selection_population(self.pop, pop_new) _, current_best = self.get_global_best_solution(pop_new) if self.compare_agent(self.g_best, current_best): pop_new[np.random.randint(0, self.pop_size)] = deepcopy(self.g_best) self.pop = pop_new