用粒子群算法解决投资组合优化问题，用python，请提供源代码和数据，代码一定要可运行

好的，下面是用粒子群算法解决投资组合优化问题的 Python 代码和数据。

数据可以使用随机生成的数据，例如：

import numpy as np

# 生成随机的收益率和方差
np.random.seed(42)
n = 10
mu = np.random.rand(n)
sigma = np.random.rand(n, n)
sigma = np.dot(sigma, sigma.T)  # 保证方差协方差矩阵为正定矩阵

代码：

import numpy as np

def objective_function(x, mu, sigma):
    # 计算投资组合的收益率和方差
    r = np.dot(mu, x)
    v = np.dot(x, np.dot(sigma, x))
    return r, v

def particle_swarm_optimization(obj_func, n_particles, n_iterations, bounds):
    # 粒子群算法
    # obj_func: 目标函数
    # n_particles: 粒子数
    # n_iterations: 迭代次数
    # bounds: 变量的取值范围

    # 初始化粒子群
    particles = np.random.rand(n_particles, len(bounds)) * (bounds[:, 1] - bounds[:, 0]) + bounds[:, 0]
    velocities = np.zeros_like(particles)
    p_best = particles.copy()
    p_best_scores = np.zeros(n_particles)
    for i in range(n_particles):
        p_best_scores[i] = obj_func(p_best[i])[0]
    g_best = p_best[p_best_scores.argmin()].copy()

    # 开始迭代
    for i in range(n_iterations):
        for j in range(n_particles):
            # 更新速度和位置
            r1, r2 = np.random.rand(2, len(bounds))
            velocities[j] = velocities[j] + r1 * (p_best[j] - particles[j]) + r2 * (g_best - particles[j])
            particles[j] = particles[j] + velocities[j]

            # 边界处理
            particles[j] = np.maximum(particles[j], bounds[:, 0])
            particles[j] = np.minimum(particles[j], bounds[:, 1])

            # 更新个体最优和全局最优
            score = obj_func(particles[j])[0]
            if score < p_best_scores[j]:
                p_best[j] = particles[j].copy()
                p_best_scores[j] = score
            if score < obj_func(g_best)[0]:
                g_best = particles[j].copy()

    return g_best

# 测试代码
np.random.seed(42)
n = 10
mu = np.random.rand(n)
sigma = np.random.rand(n, n)
sigma = np.dot(sigma, sigma.T)

bounds = np.ones((n, 2)) * np.array([0, 1])  # 变量的取值范围为 [0, 1]
x_best = particle_swarm_optimization(lambda x: objective_function(x, mu, sigma), 100, 100, bounds)
print(x_best)
print(objective_function(x_best, mu, sigma))

注意：由于粒子群算法是随机的，所以每次运行结果可能会有所不同。