How to save specific solutions based on conditions each generation

[maxgombair]

Hi, I would like to save solutions at each generation based on a selection criteria that is different from the fitness function.
The goal of this process is to study some specific solutions, it has nothing to do with parents selection.

I know that there is the save_solutions argument but its take to much time to execute the model while saving all the solutions.

So my question is:
is there a way to filter between each generation the solutions i'm interrested and save them in an array ?

I've fought about using the on_generation argument but i'm struggling to implement this, bc it needs somme kind of global variable.

[ahmedfgad]

Here is an example but it uses a global variable. It creates a list named filtered_solutions which only filters the solutions that have a fitness between 10 and 20, exclusive.

import pygad
import numpy

function_inputs = [4,-2,3.5,5,-11,-4.7]
desired_output = 44

def fitness_func(ga_instance, solution, solution_idx):
    output = numpy.sum(solution*function_inputs)
    fitness = 1.0 / (numpy.abs(output - desired_output) + 0.000001)
    return fitness

filtered_solutions = []
def on_generation(ga_instance):
    global filtered_solutions
    print(f"Generation = {ga_instance.generations_completed}")

    for idx in range(ga_instance.population.shape[0]):
        solution = ga_instance.population[idx]
        fitness = ga_instance.last_generation_fitness[idx]
        if 10 < fitness < 20:
            filtered_solutions.append(solution)

ga_instance = pygad.GA(num_generations=50,
                       num_parents_mating=5,
                       sol_per_pop=10,
                       num_genes=len(function_inputs),
                       fitness_func=fitness_func,
                       on_generation=on_generation,
                       suppress_warnings=True)

ga_instance.run()