Hello World string evolution using genetic algorithms

05-genetic-algorithms/python/string_genetic_algorithm.py

@@ -0,0 +1,85 @@
+import random
+import string
+from collections import namedtuple
+
+
+target = "Hello World"
+MAX_LEN = 50
+POPULATION_SIZE = 200
+Candidate = namedtuple("Candidate", "string fitness")
+
+
+def random_string():
+    ret_val = ""
+    for i in range(random.randint(1, MAX_LEN)):
+        ret_val += random.choice(string.ascii_letters + " ")
+    return ret_val
+
+
+def calc_fitness(candidate_string):
+    fitness = 0
+    for index, target_char in enumerate(target):
+        try:
+            if candidate_string[index] != target_char:
+                fitness -= 1
+        except IndexError:
+            fitness -= 2
+        else:
+            if len(candidate_string) > len(target):
+                fitness -= 2 * (len(candidate_string) - len(target))
+
+    return fitness
+
+
+def crossover(parent1, parent2):
+    child = parent1[:(len(parent1)/2)] + parent2[len(parent2)/2:]
+    return child
+
+
+def mutation(dna):
+    mutation_pt = random.choice(range(len(dna)))
+    dna = dna[:mutation_pt] + random.choice(string.ascii_letters + " ") + dna[mutation_pt+1:]
+    return dna
+
+
+def population(size):
+    candidates = []
+    for i in range(0, size):
+        cand_string = random_string()
+        fitness = calc_fitness(cand_string)
+        candidates.append(Candidate(string=cand_string, fitness=fitness))
+
+    return candidates
+
+
+def selection(candidates):
+    sorted_candidate_list = sorted(candidates, key=lambda candidate: candidate.fitness)
+    return sorted_candidate_list[len(sorted_candidate_list)/2:]
+
+
+def breeding(candidates):
+    breeding_pool = selection(candidates)
+    breeding_pool_copy = list(breeding_pool)
+    random.shuffle(breeding_pool)
+    cand_pairs = zip(breeding_pool, breeding_pool_copy)
+    for cand1, cand2 in cand_pairs:
+        child = crossover(cand1.string, cand2.string)
+        child = mutation(child)
+        child_fitness = calc_fitness(child)
+        breeding_pool.append(Candidate(string=child, fitness=child_fitness))
+    return breeding_pool
+
+
+def main():
+    candidates = population(POPULATION_SIZE)
+    print candidates
+    while True:
+        candidates = breeding(candidates)
+        candidates = sorted(candidates, key=lambda candidate: candidate.fitness)
+        print("Best candidate is: " + candidates[-1].string + " with fitness: %d" % candidates[-1].fitness)
+        if candidates[-1].fitness == 0:
+            break
+
+
+if __name__ == "__main__":
+    main()