粒子群算法拟合曲线python代码

好的，关于粒子群算法拟合曲线的Python代码，我可以提供以下示例代码：

import numpy as np
import matplotlib.pyplot as plt

# 定义适应度函数
def fitness(x):
    return np.sin(x) / x 

# 计算粒子群的适应度值
def pso(positions):
    # 初始化粒子位置和速度
    w = 0.7  # 惯性因子
    c1 = 1.5  # 学习因子
    c2 = 1.5

    n_particles = positions.shape[0]
    position = np.copy(positions)
    velocity = np.zeros_like(position)
    pbest_position = np.copy(position)
    pbest_fitness = np.zeros(n_particles)

    for i in range(n_particles):
        fitness_value = fitness(position[i])
        pbest_fitness[i] = fitness_value

    gbest_fitness = np.max(pbest_fitness)
    gbest_position = np.copy(pbest_position[np.argmax(pbest_fitness), :])

    error = 1e-6 # 终止条件

    while gbest_fitness > error:
        # 更新速度和位置
        for i in range(n_particles):
            r1 = np.random.rand(position.shape[1])
            r2 = np.random.rand(position.shape[1])
            velocity[i, :] = w * velocity[i, :] + \
                c1 * r1 * (pbest_position[i, :] - position[i, :]) + \
                c2 * r2 * (gbest_position - position[i, :])
            position[i, :] = position[i, :] + velocity[i, :]

            # 边界处理
            position[i, position[i, :] < -10] = -10
            position[i, position[i, :] > 10] = 10

            # 更新粒子的历史最佳位置和适应度
            fitness_value = fitness(position[i, :])
            if fitness_value > pbest_fitness[i]:
                pbest_fitness[i] = fitness_value
                pbest_position[i, :] = position[i, :]

        # 更新群体最佳位置
        if np.max(pbest_fitness) > gbest_fitness:
            gbest_fitness = np.max(pbest_fitness)
            gbest_position = np.copy(pbest_position[np.argmax(pbest_fitness), :])

    return gbest_fitness, gbest_position

particles = np.random.uniform(-10, 10, (100, 1))
fitness_value, position = pso(particles)

plt.plot(np.linspace(-10, 10, 100), fitness(np.linspace(-10, 10, 100)), label='sin(x) / x')
plt.plot(position, fitness(position), 'ro', label='PSO')
plt.legend()
plt.show()

以上代码是一个简单的粒子群算法拟合曲线的示例代码，代码中包含对适应度函数的定义、粒子群的初始化、速度和位置的更新、边界处理、以及群体最佳位置的更新等实现。需要注意的是，这里的适应度函数是 sin(x)/x，你可以根据具体需求修改适应度函数的定义。