0

To resolve a quadratic minimum problem I used scipy.Optimize.Minimize.

First I initialize my function and some points :

def func(x):
    return 1/4*(x[0]**2+x[1]**2) + x[2]

P1=[]
for i in range(3):
    P1.append((rd.uniform(-1,1), rd.uniform(-1,1)))

P2=[]
for i in range(3):
    P2.append((rd.uniform(-1,1), rd.uniform(-1,1)))

But then I have a problem with the contraints. When I do it there is no error, it works :

cons = [{'type': 'ineq', 'fun': lambda x: x[2] - x[0] * P1[0][0] - x[1] * P1[0][1] - (P1[0][0] ** 2 + P1[0][1] ** 2)},
        {'type': 'ineq', 'fun': lambda x: x[2] - x[0] * P1[1][0] - x[1] * P1[1][1] - (P1[1][0] ** 2 + P1[1][1] ** 2)},
        {'type': 'ineq', 'fun': lambda x: x[2] - x[0] * P1[2][0] - x[1] * P1[2][1] - (P1[2][0] ** 2 + P1[2][1] ** 2)}]

for i in range(len(P2)):
    cons.append({'type': 'ineq', 'fun': lambda x: -x[2] + x[0]*P2[i][0] + x[1]*P2[i][1] + (P2[i][0]**2+P2[i][1]**2)})

But when I do it, it return RuntimeWarning: invalid value encountered in double_scalars and doesn't work :

cons=[]
for i in range(0, len(P1)):
    cons.append({'type': 'ineq', 'fun': lambda x: x[2] - x[0] * P1[i][0] - x[1] * P1[i][1] - (P1[i][0] ** 2 + P1[i][1] ** 2)})

for i in range(len(P2)):
    cons.append({'type': 'ineq', 'fun': lambda x: -x[2] + x[0]*P2[i][0] + x[1]*P2[i][1] + (P2[i][0]**2+P2[i][1]**2)})

PS : Here is the end :

x0 = np.array([0, 0, 0])
res = minimize(func, x0, constraints=cons)

I don't know where is my mistake in the case that doesn't work. Thank you in advance :)

Vishnu Dasu
  • 533
  • 5
  • 18
Raphael Monin
  • 13
  • 2
  • I found the solution. Here is the answer : https://stackoverflow.com/questions/45491376/scipy-optimization-not-running-when-i-try-to-set-constraints-using-a-for-loop/45493887#45493887 – Raphael Monin Jun 06 '19 at 12:26
  • Possible duplicate of [Scipy Optimization Not Running when I try to set constraints using a for loop](https://stackoverflow.com/questions/45491376/scipy-optimization-not-running-when-i-try-to-set-constraints-using-a-for-loop) – SuperKogito Jun 06 '19 at 13:12

0 Answers0