PSO算法代码

以下是基于Python的PSO算法代码示例：

import random

class PSO:
    def __init__(self, objective_function, num_particles, num_iterations, bounds):
        self.objective_function = objective_function
        self.num_particles = num_particles
        self.num_iterations = num_iterations
        self.bounds = bounds

        self.particles = []
        for i in range(num_particles):
            particle = {}
            for j in range(len(bounds)):
                particle[j] = random.uniform(bounds[j][0], bounds[j][1])
            particle['velocity'] = [random.uniform(-1, 1) for _ in range(len(bounds))]
            particle['best_position'] = particle.copy()
            self.particles.append(particle)

        self.global_best_position = None

    def run(self):
        for i in range(self.num_iterations):
            for particle in self.particles:
                fitness = self.objective_function(*[particle[i] for i in range(len(self.bounds))])
                if particle['best_position'] is None or fitness < self.objective_function(*[particle['best_position'][i] for i in range(len(self.bounds))]):
                    particle['best_position'] = particle.copy()
                if self.global_best_position is None or fitness < self.objective_function(*[self.global_best_position[i] for i in range(len(self.bounds))]):
                    self.global_best_position = particle.copy()

            for particle in self.particles:
                for j in range(len(self.bounds)):
                    v = particle['velocity'][j] + 2 * random.random() * (particle['best_position'][j] - particle[j]) + 2 * random.random() * (self.global_best_position[j] - particle[j])
                    particle['velocity'][j] = min(1, max(-1, v))
                    particle[j] = particle[j] + particle['velocity'][j]

                    if particle[j] < self.bounds[j][0]:
                        particle[j] = self.bounds[j][0]
                    if particle[j] > self.bounds[j][1]:
                        particle[j] = self.bounds[j][1]

        return self.global_best_position

其中，objective_function表示优化的目标函数，num_particles为粒子群的数量，num_iterations为迭代次数，bounds为变量的取值范围。在运行PSO算法时，可以调用该类的run方法，返回全局最优解的位置。