Source code for mealpy.human_based.CHIO

#!/usr/bin/env python
# Created by "Thieu" at 19:24, 09/05/2020 ----------%
#       Email: nguyenthieu2102@gmail.com            %
#       Github: https://github.com/thieu1995        %
# --------------------------------------------------%

import numpy as np
from mealpy.optimizer import Optimizer
from mealpy.utils.target import Target


[docs]class OriginalCHIO(Optimizer): """ The original version of: Coronavirus Herd Immunity Optimization (CHIO) Links: 1. https://link.springer.com/article/10.1007/s00521-020-05296-6 Hyper-parameters should fine-tune in approximate range to get faster convergence toward the global optimum: + brr (float): [0.05, 0.2], Basic reproduction rate, default=0.15 + max_age (int): [5, 20], Maximum infected cases age, default=10 Examples ~~~~~~~~ >>> import numpy as np >>> from mealpy import FloatVar, CHIO >>> >>> 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 = CHIO.OriginalCHIO(epoch=1000, pop_size=50, brr = 0.15, max_age = 10) >>> 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] Al-Betar, M.A., Alyasseri, Z.A.A., Awadallah, M.A. et al. Coronavirus herd immunity optimizer (CHIO). Neural Comput & Applic 33, 5011–5042 (2021). https://doi.org/10.1007/s00521-020-05296-6 """ def __init__(self, epoch: int = 10000, pop_size: int = 100, brr: float = 0.15, max_age: int = 10, **kwargs: object) -> None: """ Args: epoch (int): maximum number of iterations, default = 10000 pop_size (int): number of population size, default = 100 brr (float): Basic reproduction rate, default=0.15 max_age (int): Maximum infected cases age, default=10 """ 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.brr = self.validator.check_float("brr", brr, (0, 1.0)) self.max_age = self.validator.check_int("max_age", max_age, [1, 1+int(epoch/5)]) self.set_parameters(["epoch", "pop_size", "brr", "max_age"])
[docs] def initialize_variables(self): self.immunity_type_list = self.generator.integers(0, 3, self.pop_size) # Randint [0, 1, 2] self.age_list = np.zeros(self.pop_size) # Control the age of each position self.finished = False
[docs] def evolve(self, epoch): """ The main operations (equations) of algorithm. Inherit from Optimizer class Args: epoch (int): The current iteration """ pop_new = [] is_corona_list = [False, ] * self.pop_size for i in range(0, self.pop_size): pos_new = self.pop[i].solution.copy() for j in range(0, self.problem.n_dims): rand = self.generator.uniform() if rand < (1.0 / 3) * self.brr: idx_candidates = np.where(self.immunity_type_list == 1) # Infected list if idx_candidates[0].size == 0: self.finished = True # print("Epoch: {}, i: {}, immunity_list: {}".format(epoch, i, self.immunity_type_list)) break idx_selected = self.generator.choice(idx_candidates[0]) pos_new[j] = self.pop[i].solution[j] + self.generator.uniform() * (self.pop[i].solution[j] - self.pop[idx_selected].solution[j]) is_corona_list[i] = True elif (1.0 / 3) * self.brr <= rand < (2.0 / 3) * self.brr: idx_candidates = np.where(self.immunity_type_list == 0) # Susceptible list idx_selected = self.generator.choice(idx_candidates[0]) pos_new[j] = self.pop[i].solution[j] + self.generator.uniform() * (self.pop[i].solution[j] - self.pop[idx_selected].solution[j]) elif (2.0 / 3) * self.brr <= rand < self.brr: idx_candidates = np.where(self.immunity_type_list == 2) # Immunity list fit_list = np.array([self.pop[item].target.fitness for item in idx_candidates[0]]) idx_selected = idx_candidates[0][np.argmin(fit_list)] # Found the index of best fitness pos_new[j] = self.pop[i].solution[j] + self.generator.uniform() * (self.pop[i].solution[j] - self.pop[idx_selected].solution[j]) if self.finished: break pos_new = self.correct_solution(pos_new) agent = self.generate_empty_agent(pos_new) pop_new.append(agent) if self.mode not in self.AVAILABLE_MODES: pop_new[-1].target = self.get_target(pos_new) pop_new = self.update_target_for_population(pop_new) if len(pop_new) != self.pop_size: pop_child = self.generate_population(self.pop_size - len(pop_new)) pop_new = pop_new + pop_child for idx in range(0, self.pop_size): # Step 4: Update herd immunity population if self.compare_target(pop_new[idx].target, self.pop[idx].target, self.problem.minmax): self.pop[idx] = pop_new[idx].copy() else: self.age_list[idx] += 1 ## Calculate immunity mean of population fit_list = np.array([agent.target.fitness for agent in self.pop]) delta_fx = np.mean(fit_list) if self.compare_fitness(pop_new[idx].target.fitness, delta_fx, self.problem.minmax) and self.immunity_type_list[idx] == 0 and is_corona_list[idx]: self.immunity_type_list[idx] = 1 self.age_list[idx] = 1 if self.compare_fitness(delta_fx, pop_new[idx].target.fitness, self.problem.minmax) and (self.immunity_type_list[idx] == 1): self.immunity_type_list[idx] = 2 self.age_list[idx] = 0 # Step 5: Fatality condition if (self.age_list[idx] >= self.max_age) and (self.immunity_type_list[idx] == 1): self.pop[idx] = self.generate_agent() self.immunity_type_list[idx] = 0 self.age_list[idx] = 0
[docs]class DevCHIO(OriginalCHIO): """ The developed version of: Coronavirus Herd Immunity Optimization (CHIO) Hyper-parameters should fine-tune in approximate range to get faster convergence toward the global optimum: + brr (float): [0.05, 0.2], Basic reproduction rate, default=0.15 + max_age (int): [5, 20], Maximum infected cases age, default=10 Examples ~~~~~~~~ >>> import numpy as np >>> from mealpy import FloatVar, CHIO >>> >>> 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 = CHIO.DevCHIO(epoch=1000, pop_size=50, brr = 0.15, max_age = 10) >>> 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}") """ def __init__(self, epoch: int = 10000, pop_size: int = 100, brr: float = 0.15, max_age: int = 10, **kwargs: object) -> None: """ Args: epoch (int): maximum number of iterations, default = 10000 pop_size (int): number of population size, default = 100 brr (float): Basic reproduction rate, default=0.15 max_age (int): Maximum infected cases age, default=10 """ super().__init__(epoch, pop_size, brr, max_age, **kwargs)
[docs] def evolve(self, epoch): """ The main operations (equations) of algorithm. Inherit from Optimizer class Args: epoch (int): The current iteration """ pop_new = [] is_corona_list = [False, ] * self.pop_size for i in range(0, self.pop_size): pos_new = self.pop[i].solution.copy() for j in range(0, self.problem.n_dims): rand = self.generator.uniform() if rand < (1.0 / 3) * self.brr: idx_candidates = np.where(self.immunity_type_list == 1) # Infected list if idx_candidates[0].size == 0: rand_choice = self.generator.choice(range(0, self.pop_size), int(0.33 * self.pop_size), replace=False) self.immunity_type_list[rand_choice] = 1 idx_candidates = np.where(self.immunity_type_list == 1) idx_selected = self.generator.choice(idx_candidates[0]) pos_new[j] = self.pop[i].solution[j] + self.generator.uniform() * (self.pop[i].solution[j] - self.pop[idx_selected].solution[j]) is_corona_list[i] = True elif (1.0 / 3) * self.brr <= rand < (2.0 / 3) * self.brr: idx_candidates = np.where(self.immunity_type_list == 0) # Susceptible list if idx_candidates[0].size == 0: rand_choice = self.generator.choice(range(0, self.pop_size), int(0.33 * self.pop_size), replace=False) self.immunity_type_list[rand_choice] = 0 idx_candidates = np.where(self.immunity_type_list == 0) idx_selected = self.generator.choice(idx_candidates[0]) pos_new[j] = self.pop[i].solution[j] + self.generator.uniform() * (self.pop[i].solution[j] - self.pop[idx_selected].solution[j]) elif (2.0 / 3) * self.brr <= rand < self.brr: idx_candidates = np.where(self.immunity_type_list == 2) # Immunity list fit_list = np.array([self.pop[item].target.fitness for item in idx_candidates[0]]) idx_selected = idx_candidates[0][np.argmin(fit_list)] # Found the index of best fitness pos_new[j] = self.pop[i].solution[j] + self.generator.uniform() * (self.pop[i].solution[j] - self.pop[idx_selected].solution[j]) if self.finished: break pos_new = self.correct_solution(pos_new) agent = self.generate_empty_agent(pos_new) pop_new.append(agent) if self.mode not in self.AVAILABLE_MODES: pop_new[-1].target = self.get_target(pos_new) pop_new = self.update_target_for_population(pop_new) for idx in range(0, self.pop_size): # Step 4: Update herd immunity population if self.compare_target(pop_new[idx].target, self.pop[idx].target, self.problem.minmax): self.pop[idx] = pop_new[idx].copy() else: self.age_list[idx] += 1 ## Calculate immunity mean of population fit_list = np.array([agent.target.fitness for agent in self.pop]) delta_fx = np.mean(fit_list) if self.compare_fitness(pop_new[idx].target.fitness, delta_fx, self.problem.minmax) and (self.immunity_type_list[idx] == 0) and is_corona_list[idx]: self.immunity_type_list[idx] = 1 self.age_list[idx] = 1 if self.compare_fitness(delta_fx, pop_new[idx].target.fitness, self.problem.minmax) and (self.immunity_type_list[idx] == 1): self.immunity_type_list[idx] = 2 self.age_list[idx] = 0 # Step 5: Fatality condition if (self.age_list[idx] >= self.max_age) and (self.immunity_type_list[idx] == 1): self.pop[idx] = self.generate_agent() self.immunity_type_list[idx] = 0 self.age_list[idx] = 0