from detpy.DETAlgs.base import BaseAlg
from detpy.DETAlgs.data.alg_data import AADEData
from detpy.DETAlgs.methods.methods_aade import aade_mutation, aade_crossing, aade_selection, \
aade_adapat_parameters
from detpy.models.enums.boundary_constrain import fix_boundary_constraints
[docs]
class AADE(BaseAlg):
"""
AADE
Links:
https://ieeexplore.ieee.org/document/8819749
References:
V. Sharma, S. Agarwal and P. K. Verma, "Auto Adaptive Differential Evolution Algorithm,"
2019 3rd International Conference on Computing Methodologies and Communication (ICCMC),
Erode, India, 2019, pp. 958-963, doi: 10.1109/ICCMC.2019.8819749.
"""
def __init__(self, params: AADEData, db_conn=None, db_auto_write=False):
super().__init__(AADE.__name__, params, db_conn, db_auto_write)
self.mutation_factors = [[params.mutation_factor, False] for _ in range(params.population_size)]
self.crossover_rates = [[params.crossover_rate, False] for _ in range(params.population_size)]
[docs]
def next_epoch(self):
# New population after mutation
v_pop = aade_mutation(self._pop, mutation_factors=self.mutation_factors)
# Apply boundary constrains on population in place
fix_boundary_constraints(v_pop, self.boundary_constraints_fun)
# New population after crossing
u_pop = aade_crossing(self._pop, v_pop, crossover_rates=self.crossover_rates)
# Update values before selection
u_pop.update_fitness_values(self._function.eval, self.parallel_processing)
# Select new population
new_pop = aade_selection(self._pop, u_pop, self.mutation_factors, self.crossover_rates)
aade_adapat_parameters(self._pop, self.mutation_factors, self.crossover_rates)
# Override data
self._pop = new_pop
self._epoch_number += 1