Python GEKKO的二元优化

我试图通过随机定义一个NxN对称矩阵和一个N维偏差向量，用Gekko和

qobj

解决一个二次二元优化问题。然而，我观察到计算时间低得令人怀疑：用N=20求解2e-2s，用0.05s求解200维问题。此外，我从未得到超过3-4次迭代。我是不是遗漏了什么

import numpy as np
N = 20
#create square symmetric matrix for the quadratic term
b = np.random.normal(0,1,(N,N))
Q = (b + b.T)/2
#bias vector for linear term
c=np.random.normal(0,1,N)
from gekko import GEKKO
m = GEKKO(remote=False)
z = m.Array(m.Var,N,integer=True,lb=0,ub=1,value=1)
m.qobj(c,A=Q,x=z,otype='min')
m.solve(disp=True)

---

任何建议都将不胜感激

切换到APOPT解算器以获得混合整数解

m.options.SOLVER=1

这是完整的脚本。MIQP问题通常非常快。如果问题是非线性（NLP）和混合整数元素（MINLP），则速度会显著减慢

将numpy导入为np
N=200
#为二次项创建平方对称矩阵
b=np.随机.正态（0,1，（N，N））
Q=（b+b.T）/2
#线性项的偏置矢量
c=np.随机.正态（0,1，N）
从gekko进口gekko
m=GEKKO（远程=False）
z=m.Array（m.Var，N，integer=True，lb=0，ub=1，value=1）
m、 qobj（c，A=Q，x=z，otype='min'）
m、 选项。解算器=1
m、 求解（disp=True）
打印（z）

---

谢谢你的回答！我想问一下，这类问题是否有500个变量的限制？这就是我观察到的…变量的数量没有限制。然而，计算时间可能会增加。

[[0.0] [1.0] [1.0] [0.0] [1.0] [1.0] [1.0] [1.0] [1.0] [0.0] [1.0] [1.0]
 [1.0] [1.0] [1.0] [1.0] [1.0] [1.0] [1.0] [1.0] [0.0] [1.0] [1.0] [1.0]
 [1.0] [1.0] [1.0] [1.0] [0.0] [1.0] [0.0] [1.0] [1.0] [0.0] [0.0] [0.0]
 [0.0] [1.0] [1.0] [0.0] [0.0] [1.0] [1.0] [1.0] [1.0] [1.0] [0.0] [1.0]
 [1.0] [1.0] [0.0] [1.0] [1.0] [1.0] [1.0] [1.0] [1.0] [1.0] [1.0] [0.0]
 [0.0] [1.0] [0.0] [0.0] [0.0] [0.0] [1.0] [1.0] [0.0] [0.0] [0.0] [0.0]
 [1.0] [1.0] [0.0] [0.0] [1.0] [1.0] [0.0] [1.0] [1.0] [0.0] [1.0] [1.0]
 [0.0] [1.0] [0.0] [0.0] [1.0] [1.0] [1.0] [1.0] [1.0] [1.0] [1.0] [1.0]
 [0.0] [1.0] [1.0] [1.0] [1.0] [1.0] [1.0] [1.0] [1.0] [1.0] [1.0] [1.0]
 [1.0] [1.0] [1.0] [0.0] [0.0] [0.0] [1.0] [1.0] [1.0] [1.0] [1.0] [1.0]
 [0.0] [0.0] [1.0] [1.0] [0.0] [1.0] [0.0] [1.0] [0.0] [1.0] [1.0] [1.0]
 [0.0] [0.0] [0.0] [0.0] [1.0] [0.0] [0.0] [1.0] [1.0] [0.0] [0.0] [1.0]
 [1.0] [1.0] [0.0] [1.0] [0.0] [0.0] [1.0] [1.0] [1.0] [1.0] [0.0] [1.0]
 [0.0] [0.0] [0.0] [0.0] [0.0] [0.0] [1.0] [1.0] [1.0] [1.0] [1.0] [0.0]
 [0.0] [0.0] [1.0] [1.0] [1.0] [1.0] [0.0] [1.0] [0.0] [0.0] [1.0] [1.0]
 [0.0] [0.0] [0.0] [1.0] [1.0] [1.0] [0.0] [1.0] [0.0] [1.0] [0.0] [0.0]
 [0.0] [1.0] [1.0] [1.0] [0.0] [1.0] [1.0] [0.0]]