""" Simple multithreaded algorithm to show how the 4 phases of a genetic algorithm works (Evaluation, Selection, Crossover and Mutation) https://en.wikipedia.org/wiki/Genetic_algorithm Author: D4rkia """ from __future__ import annotations import random # Maximum size of the population. bigger could be faster but is more memory expensive N_POPULATION = 200 # Number of elements selected in every generation for evolution the selection takes # place from the best to the worst of that generation must be smaller than N_POPULATION N_SELECTED = 50 # Probability that an element of a generation can mutate changing one of its genes this # guarantees that all genes will be used during evolution MUTATION_PROBABILITY = 0.4 # just a seed to improve randomness required by the algorithm random.seed(random.randint(0, 1000)) def basic(target: str, genes: list[str], debug: bool = True) -> tuple[int, int, str]: """ Verify that the target contains no genes besides the ones inside genes variable. >>> from string import ascii_lowercase >>> basic("doctest", ascii_lowercase, debug=False)[2] 'doctest' >>> genes = list(ascii_lowercase) >>> genes.remove("e") >>> basic("test", genes) Traceback (most recent call last): ... ValueError: ['e'] is not in genes list, evolution cannot converge >>> genes.remove("s") >>> basic("test", genes) Traceback (most recent call last): ... ValueError: ['e', 's'] is not in genes list, evolution cannot converge >>> genes.remove("t") >>> basic("test", genes) Traceback (most recent call last): ... ValueError: ['e', 's', 't'] is not in genes list, evolution cannot converge """ # Verify if N_POPULATION is bigger than N_SELECTED if N_POPULATION < N_SELECTED: raise ValueError(f"{N_POPULATION} must be bigger than {N_SELECTED}") # Verify that the target contains no genes besides the ones inside genes variable. not_in_genes_list = sorted({c for c in target if c not in genes}) if not_in_genes_list: raise ValueError( f"{not_in_genes_list} is not in genes list, evolution cannot converge" ) # Generate random starting population population = [] for _ in range(N_POPULATION): population.append("".join([random.choice(genes) for i in range(len(target))])) # Just some logs to know what the algorithms is doing generation, total_population = 0, 0 # This loop will end when we will find a perfect match for our target while True: generation += 1 total_population += len(population) # Random population created now it's time to evaluate def evaluate(item: str, main_target: str = target) -> tuple[str, float]: """ Evaluate how similar the item is with the target by just counting each char in the right position >>> evaluate("Helxo Worlx", Hello World) ["Helxo Worlx", 9] """ score = len( [g for position, g in enumerate(item) if g == main_target[position]] ) return (item, float(score)) # Adding a bit of concurrency can make everything faster, # # import concurrent.futures # population_score: list[tuple[str, float]] = [] # with concurrent.futures.ThreadPoolExecutor( # max_workers=NUM_WORKERS) as executor: # futures = {executor.submit(evaluate, item) for item in population} # concurrent.futures.wait(futures) # population_score = [item.result() for item in futures] # # but with a simple algorithm like this will probably be slower # we just need to call evaluate for every item inside population population_score = [evaluate(item) for item in population] # Check if there is a matching evolution population_score = sorted(population_score, key=lambda x: x[1], reverse=True) if population_score[0][0] == target: return (generation, total_population, population_score[0][0]) # Print the Best result every 10 generation # just to know that the algorithm is working if debug and generation % 10 == 0: print( f"\nGeneration: {generation}" f"\nTotal Population:{total_population}" f"\nBest score: {population_score[0][1]}" f"\nBest string: {population_score[0][0]}" ) # Flush the old population keeping some of the best evolutions # Keeping this avoid regression of evolution population_best = population[: int(N_POPULATION / 3)] population.clear() population.extend(population_best) # Normalize population score from 0 to 1 population_score = [ (item, score / len(target)) for item, score in population_score ] # Select, Crossover and Mutate a new population def select(parent_1: tuple[str, float]) -> list[str]: """Select the second parent and generate new population""" pop = [] # Generate more child proportionally to the fitness score child_n = int(parent_1[1] * 100) + 1 child_n = 10 if child_n >= 10 else child_n for _ in range(child_n): parent_2 = population_score[random.randint(0, N_SELECTED)][0] child_1, child_2 = crossover(parent_1[0], parent_2) # Append new string to the population list pop.append(mutate(child_1)) pop.append(mutate(child_2)) return pop def crossover(parent_1: str, parent_2: str) -> tuple[str, str]: """Slice and combine two string in a random point""" random_slice = random.randint(0, len(parent_1) - 1) child_1 = parent_1[:random_slice] + parent_2[random_slice:] child_2 = parent_2[:random_slice] + parent_1[random_slice:] return (child_1, child_2) def mutate(child: str) -> str: """Mutate a random gene of a child with another one from the list""" child_list = list(child) if random.uniform(0, 1) < MUTATION_PROBABILITY: child_list[random.randint(0, len(child)) - 1] = random.choice(genes) return "".join(child_list) # This is Selection for i in range(N_SELECTED): population.extend(select(population_score[int(i)])) # Check if the population has already reached the maximum value and if so, # break the cycle. if this check is disabled the algorithm will take # forever to compute large strings but will also calculate small string in # a lot fewer generations if len(population) > N_POPULATION: break if __name__ == "__main__": target_str = ( "This is a genetic algorithm to evaluate, combine, evolve, and mutate a string!" ) genes_list = list( " ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklm" "nopqrstuvwxyz.,;!?+-*#@^'èéòà€ù=)(&%$£/\\" ) print( "\nGeneration: %s\nTotal Population: %s\nTarget: %s" % basic(target_str, genes_list) )