python遗传算法程序

Python是一种非常适合实现遗传算法的语言，因为其语法简洁，库丰富。下面是一个基本的Python遗传算法程序框架，它会解决一个简单的函数优化问题：

import random
import numpy as np

# 定义适应度函数 (这里假设我们要最小化一个简单的一维函数)
def fitness_function(x):
    return x**2  # 示例：二次函数

# 遗传算法参数
pop_size = 50   # 种群规模
num_generations = 100  # 迭代次数
mutation_rate = 0.01  # 变异率
crossover_probability = 0.9  # 交叉概率

# 初始化种群
population = [np.random.uniform(-10, 10) for _ in range(pop_size)]

# 辅助函数：选择、交叉和变异
def selection(population, fitness):
    return random.choices(population, weights=fitness, k=len(population))

def crossover(parents):
    offspring = []
    for i in range(len(parents)//2):  # 半双倍交叉
        if random.random() < crossover_probability:
            point = random.randint(1, len(parents[0])-1)
            child1 = parents[i][:point] + parents[i+1][point:]
            child2 = parents[i+1][:point] + parents[i][point:]
        else:
            child1 = parents[i]
            child2 = parents[i+1]
        offspring.append(child1)
        offspring.append(child2)
    return offspring

def mutate(individual):
    if random.random() < mutation_rate:
        individual[random.randint(0, len(individual)-1)] += random.uniform(-0.5, 0.5)
    return individual

# 主遗传算法循环
best_solution = None
best_fitness = float('inf')
for generation in range(num_generations):
    # 计算适应度
    population_fitness = [fitness_function(ind) for ind in population]

    # 更新种群
    new_population = selection(population, population_fitness)
    new_population = crossover(new_population)
    new_population = [mutate(ind) for ind in new_population]

    # 保存当前最佳解
    current_best = min(new_population, key=fitness_function)
    if fitness_function(current_best) < best_fitness:
        best_solution = current_best
        best_fitness = fitness_function(current_best)

    population = new_population

print("Best solution found:", best_solution, "with a fitness of", best_fitness)