【Python练习】072. 编写一个函数，实现简单的蚁群算法

视睿

于 2025-08-02 08:30:00 发布

阅读量1k

点赞数 29

CC 4.0 BY-SA版权

分类专栏：从零开始学习机器人文章标签： python 算法开发语言机器人

本文链接：https://blog.csdn.net/2301_81238088/article/details/149688943

072. 编写一个函数，实现简单的蚁群算法

072. 编写一个函数，实现简单的蚁群算法

072. 编写一个函数，实现简单的蚁群算法

蚁群算法（Ant Colony Optimization, ACO）是一种模拟蚂蚁觅食行为的优化算法，常用于解决组合优化问题，如旅行商问题（TSP）。

示例

假设我们有一个旅行商问题（TSP），需要找到访问多个城市并返回起点的最短路径。

蚁群算法的基本步骤

初始化：初始化信息素浓度。
蚂蚁移动：每只蚂蚁根据信息素浓度选择路径。
更新信息素：根据蚂蚁走过的路径更新信息素浓度。
迭代：重复上述步骤，直到满足终止条件（如达到最大迭代次数）。

代码

import numpy as np
import random

# 计算城市之间的距离矩阵
def calculate_distance_matrix(cities):
    num_cities = len(cities)
    distance_matrix = np.zeros((num_cities, num_cities))
    for i in range(num_cities):
        for j in range(i + 1, num_cities):
            distance = np.linalg.norm(cities[i] - cities[j])
            distance_matrix[i][j] = distance
            distance_matrix[j][i] = distance
    return distance_matrix

# 蚁群算法
def ant_colony_optimization(cities, num_ants, num_iterations, decay, alpha=1, beta=2):
    """
    蚁群算法解决 TSP 问题

    :param cities: 城市坐标列表
    :param num_ants: 蚂蚁数量
    :param num_iterations: 迭代次数
    :param decay: 信息素衰减率
    :param alpha: 信息素重要性参数
    :param beta: 启发式信息重要性参数
    :return: 最短路径和路径长度
    """
    num_cities = len(cities)
    distance_matrix = calculate_distance_matrix(cities)
    pheromone_matrix = np.ones((num_cities, num_cities))

    best_path = None
    best_path_length = float('inf')

    for iteration in range(num_iterations):
        all_paths = []
        all_path_lengths = []

        for ant in range(num_ants):
            path = [random.randint(0, num_cities - 1)]  # 随机选择起始城市
            visited = set(path)

            while len(visited) < num_cities:
                current_city = path[-1]
                probabilities = np.zeros(num_cities)

                for next_city in range(num_cities):
                    if next_city not in visited:
                        probabilities[next_city] = (pheromone_matrix[current_city][next_city] ** alpha) * \
                                                    ((1.0 / distance_matrix[current_city][next_city]) ** beta)

                probabilities /= np.sum(probabilities)
                next_city = np.random.choice(range(num_cities), p=probabilities)
                path.append(next_city)
                visited.add(next_city)

            path_length = sum(distance_matrix[path[i]][path[i + 1