filter solutions by constraints before non-dominated sorting

[ylzhouchris]

opt-greedy/CODE/greedy/evolution.py

Lines 102 to 117 in 3c11ad7

	self.utils.fast_nondominated_sort(neighbors)
	cnt = 0
	for front in neighbors.fronts:
	self.utils.calculate_crowding_distance(front)
	cnt += 1
	#------分析front 0中的解，满足constraint的保留到final solution，并去掉。--------
	to_remove = Population()
	to_remove.extend(ind for ind in neighbors if ind.constraint[0] == self.max_sensor)
	population_final.extend(ind for ind in neighbors.fronts[0] if ind.constraint[0] == self.max_sensor)
	# print('population_final:',len(population_final))
	for ind in to_remove:
	neighbors.remove(ind)

Filter solutions through constraints before non-dominated sorting is more efficient.