ga遗传算法Python代码

GA（Genetic Algorithm，遗传算法）是一种基于自然选择和遗传机制的优化搜索算法，常用于解决复杂的优化问题。在Python中实现GA的基本步骤包括初始化种群、适应度函数评估、选择操作、交叉（Crossover）、变异（Mutation）以及迭代直到达到终止条件。

下面是一个简单的GA基本框架的Python代码示例：

import numpy as np

# 初始化种群
def initialize_population(size, dimensions):
    return np.random.rand(size, dimensions)

# 计算适应度函数
def fitness_function(individuals):
    # 这里假设适应度函数是一个简单的求和，实际应用中根据目标函数调整
    return sum(individuals**2, axis=1)

# 遗传操作
def selection(population, fitness):
    parents = np.random.choice(population, size=len(population), replace=True, p=fitness / fitness.sum())
    return parents

def crossover(parents, probability):
    offspring = []
    for _ in range(len(parents)):
        if np.random.rand() < probability:
            # 使用单点交叉
            cut_point = np.random.randint(0, len(parents[0]))
            offspring.append(np.concatenate((parents[0][:cut_point], parents[1][cut_point:])))
        else:
            offspring.append(parents[np.random.randint(0, 2)])
    return offspring

def mutation(offspring, mutation_rate):
    mutated_offspring = offspring.copy()
    for individual in mutated_offspring:
        for i in range(len(individual)):
            if np.random.rand() < mutation_rate:
                individual[i] += np.random.randn() * 0.1  # 可根据需要调整变异范围
    return mutated_offspring

def ga_algorithm(pop_size, generations, dimensions, crossover_prob, mutation_prob):
    population = initialize_population(pop_size, dimensions)
    for generation in range(generations):
        fitness_values = fitness_function(population)
        selected_pop = selection(population, fitness_values)
        offspring = crossover(selected_pop, crossover_prob)
        offspring = mutation(offspring, mutation_prob)
        # 如果有早停策略可以在这里检查最优解
        population = offspring
    best_solution = population[np.argmax(fitness_values)]
    return best_solution, fitness_values.max()

# 示例参数设置
pop_size = 50
generations = 100
dimensions = 10
crossover_prob = 0.8
mutation_prob = 0.1

best_solution, max_fitness = ga_algorithm(pop_size, generations, dimensions, crossover_prob, mutation_prob)
print(f"最佳解：{best_solution}, 最大适应度：{max_fitness}")