当前位置:首页 » 《随便一记》 » 正文

遗传算法模型Python代码——用Python实现遗传算法案例

3 人参与  2024年11月17日 18:02  分类 : 《随便一记》  评论

点击全文阅读


一、遗传算法概述

1.1适用范围

遗传算法(Genetic Algorithm, GA)是一种启发式搜索算法,广泛应用于以下领域:

优化问题:如函数优化、路径规划、资源分配等。机器学习:用于特征选择、超参数优化等。经济与金融:如投资组合优化、期权定价等。工程设计:如电路设计、结构优化等。

1.2步骤

遗传算法借鉴了自然界中生物进化的过程,主要包括以下几个步骤:

初始化:随机生成一组初始种群(解)。适应度评估:根据适应度函数评估每个个体的适应度。选择:根据适应度选择优秀的个体作为父代。交叉:通过交叉操作生成新的个体(子代)。变异:随机改变部分个体以增加种群多样性。替换:用子代替换父代,进入下一代迭代。

1.3优点

适用性广:可以应用于各种复杂的优化问题。全局搜索能力强:不容易陷入局部最优。易于并行化:可以利用多处理器环境加速计算。

1.4缺点

计算复杂度高:尤其是在种群规模和迭代次数较大的情况下。参数选择敏感:如交叉率、变异率等参数需要精心调整。收敛速度慢:在某些问题上可能需要较多代数才能找到满意的解。

二、Python代码示例

2.1代码示例

以下是一个简单的遗传算法示例,用于解决函数优化问题:

import numpy as npimport random# 目标函数(待优化)def objective_function(x):    return -x**2 + 4*x + 10# 初始化种群def initialize_population(size, bounds):    population = []    for _ in range(size):        individual = random.uniform(bounds[0], bounds[1])        population.append(individual)    return population# 适应度评估def evaluate_population(population):    return [objective_function(individual) for individual in population]# 选择(轮盘赌选择)def select(population, fitness):    total_fitness = sum(fitness)    probabilities = [f/total_fitness for f in fitness]    selected = np.random.choice(population, size=len(population), p=probabilities)    return selected# 交叉(单点交叉)def crossover(parent1, parent2):    point = random.randint(1, len(parent1)-1)    child1 = parent1[:point] + parent2[point:]    child2 = parent2[:point] + parent1[point:]    return child1, child2# 变异def mutate(individual, mutation_rate, bounds):    if random.random() < mutation_rate:        individual = random.uniform(bounds[0], bounds[1])    return individual# 遗传算法主函数def genetic_algorithm(objective_function, bounds, population_size, generations, mutation_rate):    # 初始化种群    population = initialize_population(population_size, bounds)        for generation in range(generations):        # 评估种群适应度        fitness = evaluate_population(population)                # 选择        selected_population = select(population, fitness)                # 生成新种群        new_population = []        for i in range(0, len(selected_population), 2):            parent1 = selected_population[i]            parent2 = selected_population[i+1]            child1, child2 = crossover(parent1, parent2)            new_population.append(mutate(child1, mutation_rate, bounds))            new_population.append(mutate(child2, mutation_rate, bounds))                population = new_population        # 返回最佳个体    best_individual = max(population, key=objective_function)    return best_individual# 参数设置bounds = [0, 10]population_size = 20generations = 50mutation_rate = 0.1# 执行遗传算法best_solution = genetic_algorithm(objective_function, bounds, population_size, generations, mutation_rate)print(f"最佳解:{best_solution}, 目标函数值:{objective_function(best_solution)}")

2.2代码解释

objective_function: 目标函数,用于评估个体的适应度。initialize_population: 初始化种群,生成一定范围内的随机个体。evaluate_population: 评估种群中每个个体的适应度。select: 轮盘赌选择,基于适应度概率选择个体。crossover: 单点交叉操作,生成子代个体。mutate: 变异操作,随机改变个体。genetic_algorithm: 遗传算法的主函数,执行初始化、选择、交叉、变异等操作。

三、可运行案例:用Python实现遗传算法

3.1案例代码

为了演示遗传算法的运行结果,我们以优化一个简单的二次函数为例,假设我们要找到函数 f(x)=-x^{2}+4x+10 的最大值。以下是详细的代码和运行结果的分析:

import numpy as npimport randomimport matplotlib.pyplot as plt# 目标函数(待优化)def objective_function(x):    return -x**2 + 4*x + 10# 初始化种群def initialize_population(size, bounds):    population = []    for _ in range(size):        individual = random.uniform(bounds[0], bounds[1])        population.append(individual)    return population# 适应度评估def evaluate_population(population):    return [objective_function(individual) for individual in population]# 选择(轮盘赌选择)def select(population, fitness):    min_fitness = min(fitness)    if min_fitness < 0:        fitness = [f - min_fitness for f in fitness]  # 平移使所有适应度非负    total_fitness = sum(fitness)    probabilities = [f/total_fitness for f in fitness]    selected = np.random.choice(population, size=len(population), p=probabilities)    return selected# 交叉(单点交叉)def crossover(parent1, parent2):    child1 = (parent1 + parent2) / 2    child2 = (parent1 + parent2) / 2    return child1, child2# 变异def mutate(individual, mutation_rate, bounds):    if random.random() < mutation_rate:        individual = random.uniform(bounds[0], bounds[1])    return individual# 遗传算法主函数def genetic_algorithm(objective_function, bounds, population_size, generations, mutation_rate):    # 初始化种群    population = initialize_population(population_size, bounds)    best_fitness_over_time = []        for generation in range(generations):        # 评估种群适应度        fitness = evaluate_population(population)        best_fitness_over_time.append(max(fitness))                # 选择        selected_population = select(population, fitness)                # 生成新种群        new_population = []        for i in range(0, len(selected_population), 2):            parent1 = selected_population[i]            parent2 = selected_population[i+1]            child1, child2 = crossover(parent1, parent2)            new_population.append(mutate(child1, mutation_rate, bounds))            new_population.append(mutate(child2, mutation_rate, bounds))                population = new_population        # 返回最佳个体及其适应度随时间变化    best_individual = max(population, key=objective_function)    return best_individual, best_fitness_over_time# 参数设置bounds = [0, 10]population_size = 20generations = 50mutation_rate = 0.1# 执行遗传算法best_solution, fitness_over_time = genetic_algorithm(objective_function, bounds, population_size, generations, mutation_rate)# 输出最佳解best_solution_value = objective_function(best_solution)# 绘制适应度变化图plt.plot(fitness_over_time)plt.xlabel('Generation')plt.ylabel('Best Fitness')plt.title('Genetic Algorithm Optimization')plt.show()best_solution, best_solution_value

3.2运行结果

Result(2.312058920926045, 13.902619229870472)

(1)运行结果

最佳解:x≈2.31x \approx 2.31x≈2.31目标函数值:f(x)≈13.90f(x) \approx 13.90f(x)≈13.90

(2)适应度变化图

适应度变化图如下所示:

3.3结果分析

最佳解与目标函数值

遗传算法找到了函数f(x)=-x^{2}+4x+10 的最大值点x\approx 2.31。目标函数值在该点的值为f(x)\approx 13.90,这是该函数的最大值。

适应度变化图

图中显示了每一代的最佳适应度值,适应度值逐渐上升并趋于稳定。适应度值的上升表示种群中个体质量的提升,算法逐步逼近最优解。

通过上述步骤,遗传算法成功找到了优化问题的近似最优解,并且从适应度变化图可以看到算法的收敛过程。如果对最终结果不满意,可以尝试调整参数(如种群规模、变异率等)以获得更好的结果。


点击全文阅读


本文链接:http://m.zhangshiyu.com/post/187825.html

<< 上一篇 下一篇 >>

  • 评论(0)
  • 赞助本站

◎欢迎参与讨论,请在这里发表您的看法、交流您的观点。

关于我们 | 我要投稿 | 免责申明

Copyright © 2020-2022 ZhangShiYu.com Rights Reserved.豫ICP备2022013469号-1