为循环存储数据

Question

我正在使用蒙特卡洛方法进行数值积分。我已经设法使函数工作，但现在我正试图将结果存储在一个矩阵中，这样我就可以将所有内容绘制在一起。

这是我的错误： ``Traceback（最近一次通话最后）： 文件“mcmc.py”，第 41 行，在 result_array = 列表（fnArea） TypeError: 'float' 对象不可迭代

这是我的代码：

from math import *
from random import *
import numpy as np

def f(a):
    return sin(a)

amin = 0.0
amax = 2.0*pi

numRep = 100  #quanto maior, mais preciso porem mais devagar
bmin = f(amin)
bmax = bmin
for i in range(numRep):
    a = amin + (amax - amin) * float(i)/numRep
    b = f(a)
    if b < bmin: bmin = b
    if b > bmax: bmax = b

retArea = (amax - amin)*(bmax - bmin) #definicao da integral
numPontos = 100
soma = 0
for j in range(numPontos):
   a = amin + (amax - amin) * random()
   b = bmin + (bmax - bmin) * random()
   if fabs(b) <= fabs(f(a)):
        if f(a) > 0 and b > 0 and b <= f(a):
            soma += 1
        if f(a) < 0 and b < 0 and b >= f(a):
            soma -= 1

fnArea = retArea * float(soma)/numPontos 
print (str(fnArea))

result_array = np.empty((0,100))
count = 100
for i in range(count):
    fnArea = retArea * (soma)/numPontos
    result_array = list(fnArea)


print (result_array)

Answer 1

您收到错误是因为

result_array = list(fnArea)

这里你试图将一个浮点数变成一个无效的列表。

这是一个工作代码，但我不确定你得到的结果是否正确。

这里我们将 result_array 设为一维零数组，然后根据 i 替换每一列。

from math import *
from random import *
import numpy as np

def f(a):
    return sin(a)

amin = 0.0
amax = 2.0*pi

numRep = 100  #quanto maior, mais preciso porem mais devagar
bmin = f(amin)
bmax = bmin
for i in range(numRep):
    a = amin + (amax - amin) * float(i)/numRep
    b = f(a)
    if b < bmin: bmin = b
    if b > bmax: bmax = b

retArea = (amax - amin)*(bmax - bmin) #definicao da integral
numPontos = 100
soma = 0
for j in range(numPontos):
   a = amin + (amax - amin) * random()
   b = bmin + (bmax - bmin) * random()
   if fabs(b) <= fabs(f(a)):
        if f(a) > 0 and b > 0 and b <= f(a):
            soma += 1
        if f(a) < 0 and b < 0 and b >= f(a):
            soma -= 1

fnArea = retArea * float(soma)/numPontos 
print (str(fnArea))

result_array = np.zeros((1,100))
count = 100
for i in range(count):
    fnArea = retArea * (soma)/numPontos
    result_array[0,i] = fnArea


print (result_array)

Answer 2

您收到此错误是因为您试图将 float object [number] 转换为 list [which is iterable]。

如果您尝试创建一个具有fnArea 值的列表列表。我已将您的 result_array 声明修改为 python 列表，而不是 numpy 数组并在 for 循环中附加 fnArea 的值。请看下面的答案：

from math import *
from random import *
import numpy as np

def f(a):
    return sin(a)

amin = 0.0
amax = 2.0*pi

numRep = 100  #quanto maior, mais preciso porem mais devagar
bmin = f(amin)
bmax = bmin
for i in range(numRep):
    a = amin + (amax - amin) * float(i)/numRep
    b = f(a)
    if b < bmin: bmin = b
    if b > bmax: bmax = b

retArea = (amax - amin)*(bmax - bmin) #definicao da integral
numPontos = 100
soma = 0
for j in range(numPontos):
   a = amin + (amax - amin) * random()
   b = bmin + (bmax - bmin) * random()
   if fabs(b) <= fabs(f(a)):
        if f(a) > 0 and b > 0 and b <= f(a):
            soma += 1
        if f(a) < 0 and b < 0 and b >= f(a):
            soma -= 1

fnArea = retArea * float(soma)/numPontos 
print (str(fnArea))

result_array = []
count = 100
for i in range(count):
    fnArea = retArea * (soma)/numPontos
    print fnArea
    result_array.append([fnArea])



print (result_array)