使用 Scipy 进行 3D 优化的界限

编程问答 2022-04-29

问题描述

我找不到答案。我想用成本函数优化一个三维数组。数组的值不得低于 1/9999 且不得高于 1.0。

例如我的数组是这样的：

myarray = np.array([[0.1,0.2,0.3],[0.1,0.3]],[[0.1,0.3]])

我这样调用优化函数：

bounds = [(1/9999,1.0) for i in myarray for j in i for k in j]
opt.minimize(fun=CostFunction,x0=myarray.flatten(),method='Powell',bounds=bounds)

但是我收到了这个错误：

Traceback (most recent call last):
  File "C:/Users/Name/PycharmProjects/OptimalisatiePreset/Main.py",line 74,in <module>
    result = opt.minimize(fun=CostFunction,bounds=bounds)
  File "C:\Users\Name\PycharmProjects\OptimalisatiePreset\venv\lib\site-packages\scipy\optimize\_minimize.py",line 610,in minimize
    return _minimize_powell(fun,x0,args,callback,bounds,**options)
  File "C:\Users\Name\PycharmProjects\OptimalisatiePreset\venv\lib\site-packages\scipy\optimize\optimize.py",line 2965,in _minimize_powell
    fval,x,direc1 = _linesearch_powell(func,direc1,File "C:\Users\Name\PycharmProjects\OptimalisatiePreset\venv\lib\site-packages\scipy\optimize\optimize.py",line 2700,in _linesearch_powell
    bound = _line_for_search(p,xi,lower_bound,upper_bound)
  File "C:\Users\Name\PycharmProjects\OptimalisatiePreset\venv\lib\site-packages\scipy\optimize\optimize.py",line 2643,in _line_for_search
    lower_bound,upper_bound = lower_bound[nonzero],upper_bound[nonzero]
IndexError: index 0 is out of bounds for axis 0 with size 0

有人知道我做错了什么吗？

解决方法

不确定您的问题是什么，以下对我有用，没有错误。答案并不完美：我认为每个 x 都应该是 1.0/9999.0，但事实并非如此。但是它似乎可以运行，也许它可以成为您的起点。

import numpy as np
from scipy.optimize import minimize

myarray = np.asarray([[[0.1,0.2,0.3],[0.1,0.3]],[[0.1,0.3]]])

bounds = [(1/9999,1.0) for i in myarray for j in i for k in j]

def CostFunction(x):
    return np.linalg.norm(x)
    
out=minimize(fun=CostFunction,x0=myarray.flatten(),method='Powell',bounds=bounds)
boundsboundsoptimizationoptimizationpythonscipyscipy