hooke jeeves算法（模式搜索法）

最新推荐文章于 2025-05-17 15:18:37 发布

笨笨鱼小牛

最新推荐文章于 2025-05-17 15:18:37 发布

阅读量1.8k

点赞数

文章标签：算法 python 机器学习

本文链接：https://blog.csdn.net/qq_39578187/article/details/123029750

版权

# -*- coding: utf-8 -*-
"""
Spyder Editor

This is a temporary script file.
"""

import numpy as np
import matplotlib.pyplot as plt

# -----------define HJ agolrathiem para
delta = 0.5 # step size
alfa = 1    # speed rate
beta = 0.5  # decrease rate
epsilon = 0.002 # min stepsize
iniPO = np.array([6,1])     # initial point
G = 10  # 探测次数

# -----------define function
# min f(x):=(1−x1)^2+5(x2−x1^2)^2
def fun(A):
    ff = (1 - A[0]) ** 2 + 5 * (A[1] - A[0] ** 2) ** 2
    return ff

iniPO_ff = np.array([fun(iniPO)])
X = iniPO
X_ff = iniPO_ff
Y = X
Y_ff =X_ff
x = iniPO
y = iniPO

# -----------explore step
dim = iniPO.shape[0] # dimation
e = [1]*dim
E = np.diag(e)
k=0
while delta > epsilon:    
    for i in range(0,dim):
       y_new_plus = y + delta * E[i,:]
       ff = fun(y)
       ff_new = np.array([fun(y_new_plus)])
       temp = Y
       tem