Neural Network 学习5 Himmelblau函数优化实例

import os

os.environ['TF_CPP_MIN_LOG_LEVEL'] = '2'
import tensorflow as tf
from tensorflow import keras
import numpy as np
from matplotlib import pyplot as plt
from mpl_toolkits.mplot3d import Axes3D


def himmelblau(x):  # 定义函数
    return (x[0] ** 2 + x[1] - 11) ** 2 + (x[0] + x[1] ** 2 - 7) ** 2

x = np.arange(-6, 6, 0.1) #arange 返回一个有起点有终点的固定步长的排列
y = np.arange(-6, 6, 0.1)
X, Y = np.meshgrid(x, y) #X中存放着所有点的横坐标，Y中存放着所有点的纵坐标
Z = himmelblau([X, Y])

fig = plt.figure('himmelblau') #画图
ax = fig.gca(projection='3d')
ax.plot_surface(X, Y, Z)
ax.view_init(60, -30)
ax.set_xlabel('x')
ax.set_ylabel('y')
plt.show()

x = tf.constant([-4., 0.])  #从初始点[4.,0.]开始优化，起始点不同，得到的极小值的坐标也不同

for step in range(200):  # 优化200次
    with tf.GradientTape() as tape:
        tape.watch([x])
        y = himmelblau(x)
    grads = tape.gradient(y, [x])[0]
    x -= 0.01 * grads # 梯度更新

    if step % 20 == 0:
        print('step{}:x={},f(x)={}'.format(step, x.numpy(), y.numpy()))