粒子群算法求解车间调度问题python代码

粒子群算法（Particle Swarm Optimization，PSO）是一种具有全局寻优能力的优化方法，可以应用于车间调度问题。下面是用Python实现车间调度问题的粒子群算法。

首先，定义函数以计算每个粒子的适应度，即车间调度的总加工时间：

def fitness_func(schedule, jobs):
    times = [0] * len(jobs)
    for i in range(len(schedule)):
        job = jobs[schedule[i]]
        if i == 0:
            times[i] = job[0] + job[1]
        else:
            times[i] = max(times[i-1], job[0]) + job[1]
    return max(times)

然后，实现粒子群算法：

# 初始化粒子
def init_particles(num_p, num_j):
    particles = []
    for i in range(num_p):
        particle = []
        for j in range(num_j):
            particle.append(random.randint(0, num_j-1))
        particles.append(particle)
    return particles

# 计算每个粒子的适应度
def update_fitness(particles, jobs):
    fitness = []
    for particle in particles:
        fitness.append(fitness_func(particle, jobs))
    return fitness

# 更新每个粒子的速度和位置
def update_particles(particles, best, w, c1, c2):
    for i in range(len(particles)):
        for j in range(len(particles[i])):
            r1 = random.uniform(0, 1)
            r2 = random.uniform(0, 1)
            particles[i][j] = int(particles[i][j] + w * (best[i][j] - particles[i][j]) + c1 * r1 * (global_best[j] - particles[i][j]) + c2 * r2 * (best_global[j] - particles[i][j]))
            if particles[i][j] < 0:
                particles[i][j] = 0
            elif particles[i][j] > len(jobs)-1:
                particles[i][j] = len(jobs)-1

# 计算全局最优解和每个粒子的最优解
def update_best(particles, best):
    for i in range(len(particles)):
        if fitness[i] < best[i][len(jobs)]:
            best[i] = particles[i] + [fitness[i]]
        if fitness[i] < best_global[len(jobs)]:
            best_global = particles[i] + [fitness[i]]

最后，运行粒子群算法的主函数：

if __name__ == '__main__':
    jobs = [(4, 5), (1, 3), (2, 4), (4, 2), (1, 5), (4, 2), (3, 5), (2, 1), (5, 2), (4, 4)]
    num_particles = 50
    num_generations = 100
    w = 0.9
    c1 = 2
    c2 = 2

    particles = init_particles(num_particles, len(jobs))
    fitness = update_fitness(particles, jobs)
    best, best_global = [], particles[0] + [fitness[0]]
    for i in range(len(particles)):
        best.append(particles[i] + [fitness[i]])

    for i in range(num_generations):
        update_particles(particles, best, w, c1, c2)
        fitness = update_fitness(particles, jobs)
        update_best(particles, best)
        print('Generation: {} Best Fitness: {}'.format(i, best_global[len(jobs)]))
    print('Best Schedule: {}'.format(best_global[:-1]))

在以上代码中，我们使用随机生成的作业数据完成车间调度问题的求解。输出结果包括每一代的最佳适应度和最终的最佳排程方案。随着迭代次数的增加，算法得到的最佳适应度越来越接近全局最优解，最终得到的排程方案也更加合理。