粒子群优化神经网络python代码

以下是一个简单的粒子群优化神经网络的 Python 代码示例：

import numpy as np

# 定义神经网络
class NeuralNetwork:
    def __init__(self, num_inputs, num_hidden, num_outputs):
        self.input_size = num_inputs
        self.hidden_size = num_hidden
        self.output_size = num_outputs

        # 初始化权重和偏置
        self.weights1 = np.random.randn(self.input_size, self.hidden_size)
        self.bias1 = np.random.randn(self.hidden_size)
        self.weights2 = np.random.randn(self.hidden_size, self.output_size)
        self.bias2 = np.random.randn(self.output_size)

    # 前向传播函数
    def forward(self, inputs):
        z1 = np.dot(inputs, self.weights1) + self.bias1
        a1 = np.tanh(z1)
        z2 = np.dot(a1, self.weights2) + self.bias2
        output = self.softmax(z2)
        return output

    # softmax函数
    def softmax(self, x):
        e_x = np.exp(x - np.max(x))
        return e_x / e_x.sum()

# 定义粒子群优化算法
class ParticleSwarmOptimizer:
    def __init__(self, num_particles, num_iterations, neural_network):
        self.num_particles = num_particles
        self.num_iterations = num_iterations
        self.neural_network = neural_network

        # 初始化粒子的位置和速度
        self.positions = np.random.uniform(-1, 1, size=(self.num_particles, self.get_num_weights()))
        self.velocities = np.zeros((self.num_particles, self.get_num_weights()))

        # 初始化全局最优位置
        self.global_best_position = None
        self.global_best_fitness = float('inf')

    # 计算适应度函数
    def fitness(self, weights):
        self.neural_network.set_weights(weights)
        loss = 0
        for input, target in training_data:
            output = self.neural_network.forward(input)
            loss += cross_entropy_error(output, target)
        fitness = 1 / (loss + 1e-10)
        return fitness

    # 粒子群优化算法主函数
    def optimize(self):
        for iteration in range(self.num_iterations):
            for i in range(self.num_particles):
                # 计算当前粒子的适应度和位置
                fitness = self.fitness(self.positions[i])
                if fitness > self.global_best_fitness:
                    self.global_best_fitness = fitness
                    self.global_best_position = self.positions[i]

                # 更新粒子速度和位置
                self.velocities[i] = self.update_velocity(self.velocities[i], self.positions[i], self.global_best_position)
                self.positions[i] += self.velocities[i]

    # 更新粒子速度
    def update_velocity(self, velocity, position, global_best_position):
        inertia_weight = 0.9
        cognitive_weight = 2
        social_weight = 2
        r1 = np.random.rand(self.get_num_weights())
        r2 = np.random.rand(self.get_num_weights())
        cognitive_component = cognitive_weight * r1 * (self.global_best_position - position)
        social_component = social_weight * r2 * (global_best_position - position)
        new_velocity = inertia_weight * velocity + cognitive_component + social_component
        return new_velocity

    # 获取权重数量
    def get_num_weights(self):
        return self.neural_network.input_size * self.neural_network.hidden_size + self.neural_network.hidden_size * self.neural_network.output_size

# 训练数据
training_data = [
    (np.array([0, 0]), np.array([1, 0])),
    (np.array([0, 1]), np.array([0, 1])),
    (np.array([1, 0]), np.array([0, 1])),
    (np.array([1, 1]), np.array([1, 0])),
]

# 定义神经网络
neural_network = NeuralNetwork(num_inputs=2, num_hidden=5, num_outputs=2)

# 定义粒子群优化算法
num_particles = 30
num_iterations = 100
pso = ParticleSwarmOptimizer(num_particles, num_iterations, neural_network)

# 运行粒子群优化算法
pso.optimize()

# 输出结果
weights = pso.global_best_position
neural_network.set_weights(weights)
for input, target in training_data:
    output = neural_network.forward(input)
    print("Input:", input, "Target:", target, "Output:", output)

注意，这只是一个简单的示例代码，可能需要进行更多的调整和修改才能适用于您的实际问题。