CODEWITHSUNDEEP
pythonarrayspython-2.7numpyvalueerror
curious_cosmo
curious_cosmo

Reputation: 1214

Value error with numpy arrays (shapes)

I keep getting a ValueError when working with numpy arrays and I can't figure out what's causing it, as it seems to be working correctly outside of my for loop. Here is my code:

import numpy as np

def x(t, x_0, w):
    return x_0*np.cos(w*t)
def x_prime(t, x_0, w):
    return -x_0*w*np.sin(w*t)

w = 1
x_0 = 1
h = 0.001
t = np.arange(0, 10, h)
y = np.array([[0, 0]]*len(t))

y[0] = [x_0, 0]
# The line below works correctly, but not inside my loop
print np.array([x_prime(1, x_0, w), -w**2 * x(1, x_0, w)])*h + y[0]
for i in range(1, len(t)):
    # Euler's method
    y[i] = y[i-1] + np.array([x_prime(t, x_0, w), -w**2 * x(t, x_0, w)]) * h

From the print line I get this output: [ 9.99158529e-01 -5.40302306e-04], so that seems to be working correctly. However, I'm getting this error at the y[i] line:

ValueError: operands could not be broadcast together with shapes (2,) (2,10000)

I'm not sure why, since my print statement earlier is basically doing the same thing, and y[i] should be the same shape. Does anyone know what the problem is?

Upvotes: 0

Views: 612

Answers (2)

monkeyking9528
monkeyking9528

Reputation: 126

No idea what your algorithm try to realize, but your seems to use all the t in one iteration, maybe my change can help you. import numpy as np

def x(t, x_0, w):
    return x_0*np.cos(w*t)
def x_prime(t, x_0, w):
    return -x_0*w*np.sin(w*t)

w = 1
x_0 = 1
h = 0.001
t = range(0, int(10/h))
y = np.array([[0, 0]]*len(t))

y[0,0] = x_0
y[0,1] = 0
# The line below works correctly, but not inside my loop
print(np.array([x_prime(1, x_0, w), -w**2 * x(1, x_0, w)])*h + y[0])
for i in range(1, len(t)):
    # Euler's method
    y[i] = y[i-1] + np.array([x_prime(t[i]*h, x_0, w), -w**2 * x(t[i]*h, x_0, w)]) * h

Upvotes: 0

fferri
fferri

Reputation: 18950

In the print line the first argument of x()/x_prime() is a scalar (1).

In the y[i] line you pass t instead, which is a 10000-elements array, resulting in np.array([x_prime(t, x_0, w), -w**2 * x(t, x_0, w)]) being a (2,10000) matrix, hence the ValueError.

Perhaps what you want to do is:

y[i] = y[i-1] + np.array([x_prime(t[i], x_0, w), -w**2 * x(t[i], x_0, w)]) * h

Upvotes: 1

Related Questions