A quickly made implementation of the Genetic Algorithm in python

genetic.py

import random


def generate_random_gens(n: int, size: int, items_count: int) -> list[set[int]]:
    res = []
    v = list(range(items_count))
    for _ in range(n):
        random.shuffle(v)
        res.append(set(v[:size]))
    return res


def cross(gen1: set[int], gen2: set[int]) -> set[int]:
    return set(random.sample(tuple(gen1 | gen2), len(gen1)))


def mutate(gen: set[int], items_count: int, chance: float) -> None:
    can_use = list(set(range(items_count)).difference(gen))
    for v in set(gen):
        if random.uniform(0, 1) <= chance:
            gen.remove(v)
            can_use.append(v)
            new = random.choice(can_use)
            can_use.remove(new)
            gen.add(new)


def evolve(gens: list[set[int]], items: list[int], nice_res: int) -> None:
    # Selection and crossover
    selected_count = len(gens) // 2
    gens.sort(key=lambda x: abs(nice_res - sum(map(lambda y: items[y], x))))
    for i in range(selected_count, len(gens)):
        gens[i] = cross(gens[random.randint(0, selected_count - 1)],
                        gens[random.randint(0, selected_count - 1)])
    # Mutation
    for i in range(len(gens)):
        mutate(gens[i], len(items), 0.1)


def main():
    items = [5, 1, 10, 20, 5, 30, 2, 5, 1, 1]
    nice = 13
    c = len(items)
    gens_count = 8
    gen_size = 5
    gens = generate_random_gens(gens_count, gen_size, c)
    for _ in range(10):
        evolve(gens, items, nice)
        print(*(sum(map(lambda x: items[x], inds)) for inds in gens))


if __name__ == "__main__":
    main()