# !/usr/bin/env python
# Created by "Thieu" at 14:53, 17/03/2020 ----------%
# Email: nguyenthieu2102@gmail.com %
# Github: https://github.com/thieu1995 %
# --------------------------------------------------%
import numpy as np
from mealpy.optimizer import Optimizer
[docs]class BaseGOA(Optimizer):
"""
The original version of: Grasshopper Optimization Algorithm (GOA)
Links:
1. http://dx.doi.org/10.1016/j.advengsoft.2017.01.004
2. https://www.mathworks.com/matlabcentral/fileexchange/61421-grasshopper-optimisation-algorithm-goa
Notes:
+ The matlab version above is not good, therefore
+ I added np.random.normal() component to Eq, 2.7
+ I changed the way to calculate distance between two location
Hyper-parameters should fine tuned in approximate range to get faster convergen toward the global optimum:
+ c_minmax (list): (c_min, c_max) -> ([0.00001, 0.01], [0.5, 2.0]), coefficient c, default = (0.00004, 1)
Examples
~~~~~~~~
>>> import numpy as np
>>> from mealpy.swarm_based.GOA import BaseGOA
>>>
>>> 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
>>> c_minmax = [0.00004, 1]
>>> model = BaseGOA(problem_dict1, epoch, pop_size, c_minmax)
>>> best_position, best_fitness = model.solve()
>>> print(f"Solution: {best_position}, Fitness: {best_fitness}")
References
~~~~~~~~~~
[1] Saremi, S., Mirjalili, S. and Lewis, A., 2017. Grasshopper optimisation algorithm:
theory and application. Advances in Engineering Software, 105, pp.30-47.
"""
def __init__(self, problem, epoch=10000, pop_size=100, c_minmax=(0.00004, 1), **kwargs):
"""
Args:
problem (dict): The problem dictionary
epoch (int): maximum number of iterations, default = 10000
pop_size (int): number of population size, default = 100
c_minmax (list): coefficient c, default = (0.00004, 1)
"""
super().__init__(problem, kwargs)
self.nfe_per_epoch = pop_size
self.sort_flag = False
self.epoch = epoch
self.pop_size = pop_size
self.c_minmax = c_minmax
def _s_function__(self, r_vector=None):
f = 0.5
l = 1.5
# Eq.(2.3) in the paper
return f * np.exp(-r_vector / l) - np.exp(-r_vector)
[docs] def evolve(self, epoch):
"""
The main operations (equations) of algorithm. Inherit from Optimizer class
Args:
epoch (int): The current iteration
"""
# Eq.(2.8) in the paper
c = self.c_minmax[1] - epoch * ((self.c_minmax[1] - self.c_minmax[0]) / self.epoch)
pop_new = []
for idx in range(0, self.pop_size):
S_i_total = np.zeros(self.problem.n_dims)
for j in range(0, self.pop_size):
dist = np.sqrt(np.sum((self.pop[idx][self.ID_POS] - self.pop[j][self.ID_POS]) ** 2))
r_ij_vector = (self.pop[idx][self.ID_POS] - self.pop[j][self.ID_POS]) / (dist + self.EPSILON) # xj - xi / dij in Eq.(2.7)
xj_xi = 2 + np.remainder(dist, 2) # |xjd - xid| in Eq. (2.7)
## The first part inside the big bracket in Eq. (2.7) 16 955 230 764 212 047 193 643
ran = (c / 2) * (self.problem.ub - self.problem.lb)
s_ij = ran * self._s_function__(xj_xi) * r_ij_vector
S_i_total += s_ij
x_new = c * np.random.normal() * S_i_total + self.g_best[self.ID_POS] # Eq. (2.7) in the paper
pos_new = self.amend_position(x_new)
pop_new.append([pos_new, None])
pop_new = self.update_fitness_population(pop_new)
self.pop = self.greedy_selection_population(self.pop, pop_new)