最优化最速下降法

最优化最速下降法
打出来全是nan
额，有空（也许永远没空）修复

import math
import numpy as np
from numpy.linalg import inv
from numpy.linalg.linalg import norm
from numpy.ma.core import getmask

def fun(x):
    return 100 * (x[0] ** 2 - x[1]) ** 2 + (x[0] - 1) ** 2

def gfun(x):
    # return np.matrix([[400*x[0]*(x[0]**2-x[1])+2*(x[0]-1)],[-200*(x[0]**2-x[1])]])
    return np.array([(400*x[0]*(x[0]**2-x[1])+2*(x[0]-1)), -4 * (-200*(x[0]**2-x[1]))])

def dampnm(fun, gfun,x0):
    maxk = 100
    rh0 = 0.55
    sigma = 0.4
    k = 0
    epsilon = 1e-5
    while(k<maxk):
        g = gfun(x0)
        d = -g
        if(np.linalg.norm(d)<epsilon):
            break
        m = 0
        mk = 0
        while(m<20):
            if(fun(x0+rh0**m*d)<fun(x0)+sigma**m*g.T*d):
                mk = m
                break
            m = m + 1
        x0 = x0 + rh0**mk*d
        k = k + 1
    x = x0
    val = fun(x0)
    print(x)
    print(val)

if __name__ == '__main__':
    x0 = np.array([[1], [0]])
    dampnm(x0)