我收到此错误消息：无法根据规则“安全”将数组数据从 dtype('O') 转换为 dtype('float64')

如何解决我收到此错误消息：无法根据规则“安全”将数组数据从 dtype('O') 转换为 dtype('float64')

这是我的代码。

import numpy as np
from scipy.optimize import minimize
import sympy as sp
sp.init_printing()
from sympy import *
from sympy import Symbol,Matrix

rom sympy import *
def make_Aij(m,n,a='a') :
    from sympy import Symbol,Matrix # just in case they aren't already loaded
    A = zeros(m,n)
    for i in range(0,m) :
        for j in range(0,n) :
            s = a+'_'+str(i)+str(j)
            exec  (s + "= Symbol('" + s + "')") # go look up what "exec" does!
            exec ("A[i,j] = " + s)
    return A
C = make_Aij(1,2,'c')
C

z = C[0,0]
g = C[0,0]**2-C[0,1]**2
g
h = C[0,0]
h

#Function defined
def function(h):
    return g
g

#Jacobian working with sympy
q = diff(g,C[0,0])
q

#Jacobian final
def jacobian(h):
    return q
q

Hf = diff(q,0])
Hf

#Hessianf
def Hessianf(h):
    return Hf
Hf

from scipy.optimize import Bounds
bounds = Bounds(-1,1)

h0 = (0*C[0,1])

res = minimize(function,h0,method='trust-constr',jac=jacobian,hess=Hessianf,options={'verbose': 1},bounds=bounds)

错误信息： -------------------------------------------------- -------------------------

TypeError                                 Traceback (most recent call last)
<ipython-input-41-94c28f7f32b2> in <module>
----> 1 res = minimize(function,2 
      3                options={'verbose': 1},bounds=bounds)

~\anaconda3\lib\site-packages\scipy\optimize\_minimize.py in minimize(fun,x0,args,method,jac,hess,hessp,bounds,constraints,tol,callback,options)
    626                                constraints,callback=callback,**options)
    627     elif meth == 'trust-constr':
--> 628         return _minimize_trustregion_constr(fun,629                                             bounds,630                                             callback=callback,**options)

~\anaconda3\lib\site-packages\scipy\optimize\_trustregion_constr\minimize_trustregion_constr.py in _minimize_trustregion_constr(fun,grad,xtol,gtol,barrier_tol,sparse_jacobian,maxiter,verbose,finite_diff_rel_step,initial_constr_penalty,initial_tr_radius,initial_barrier_parameter,initial_barrier_tolerance,factorization_method,disp)
    507 
    508     elif method == 'tr_interior_point':
--> 509         _,result = tr_interior_point(
    510             objective.fun,objective.grad,lagrangian_hess,511             n_vars,canonical.n_ineq,canonical.n_eq,~\anaconda3\lib\site-packages\scipy\optimize\_trustregion_constr\tr_interior_point.py in tr_interior_point(fun,lagr_hess,n_vars,n_ineq,n_eq,constr,fun0,grad0,constr_ineq0,jac_ineq0,constr_eq0,jac_eq0,stop_criteria,enforce_feasibility,state,initial_tolerance,initial_penalty,initial_trust_radius,factorization_method)
    319     while True:
    320         # Solve SQP subproblem
--> 321         z,state = equality_constrained_sqp(
    322             subprob.function_and_constraints,323             subprob.gradient_and_jacobian,~\anaconda3\lib\site-packages\scipy\optimize\_trustregion_constr\equality_constrained_sqp.py in equality_constrained_sqp(fun_and_constr,grad_and_jac,constr0,jac0,trust_lb,trust_ub,scaling)
     80     Z,LS,Y = projections(A,factorization_method)
     81     # Compute least-square lagrange multipliers
---> 82     v = -LS.dot(c)
     83     # Compute Hessian
     84     H = lagr_hess(x,v)

~\anaconda3\lib\site-packages\scipy\sparse\linalg\interface.py in dot(self,x)
    416 
    417             if x.ndim == 1 or x.ndim == 2 and x.shape[1] == 1:
--> 418                 return self.matvec(x)
    419             elif x.ndim == 2:
    420                 return self.matmat(x)

~\anaconda3\lib\site-packages\scipy\sparse\linalg\interface.py in matvec(self,x)
    230             raise ValueError('dimension mismatch')
    231 
--> 232         y = self._matvec(x)
    233 
    234         if isinstance(x,np.matrix):

~\anaconda3\lib\site-packages\scipy\sparse\linalg\interface.py in _matvec(self,x)
    528 
    529     def _matvec(self,x):
--> 530         return self.__matvec_impl(x)
    531 
    532     def _rmatvec(self,x):

~\anaconda3\lib\site-packages\scipy\optimize\_trustregion_constr\projections.py in least_squares(x)
    151         # lu_sol = [aux]
    152         #          [ z ]
--> 153         lu_sol = solve(v)
    154         # return z = inv(A A.T) A x
    155         return lu_sol[n:m+n]

TypeError: Cannot cast array data from dtype('O') to dtype('float64') according to the rule 'safe'

我只想计算沿 2 变量函数 f(c00,c01) 的 c00 轴的最小值，并有一些界限。因此，我应该得到值 c00 作为 C01 的函数（这里 c01 是一个参数）。我认为问题来自与 scipy.optimize 相关的 sympy 函数。

提前致谢。

解决方法

符号 sympy 与数字 scipy 和 numpy 不能很好地混合。数字函数不理解 sympy 的符号。

为了让事情一起工作，所有符号函数都需要转换为 numpy 等价物。 sympy 的 lambdify 可以将 sympy 表达式转换为 numpy 函数。在您的代码中，您可以按如下方式使用它：

np_function = sp.lambdify(h,function(h))
np_jacobian = sp.lambdify(h,jacobian(h))
np_hessian = sp.lambdify(h,Hessianf(h))

res = minimize(np_function,h0,method='trust-constr',jac=np_jacobian,hess=np_hessian,options={'verbose': 1},bounds=bounds)