I am writing a code to perform a Fourth Order Runge-Kutta numerical approximation with adaptive step size.
def f(x, t): #integrating function
return -x
def exact(t): #exact solution
return np.exp(-t)
def Rk4(x0, t0, dt): #Runge-Kutta Fourth Order
t = np.arange(0, 1+dt, dt)
n = len(t)
x = np.array([x0]*n)
x[0],t[0] = x0,t0
for i in range(n-1):
h = t[i+1]-t[i]
k1 = h*f(x[i], t[i])
k2 = h*f(x[i] + c21*k1, t[i] + c20*h)
k3 = h*f(x[i] + c31*k1 + c32*k2, t[i] + c30*h)
k4 = h*f(x[i] + c41*k1 + c42*k2 + c43*k3, t[i] + c40*h)
k5 = h*f(x[i] + c51*k1 + c52*k2 + c53*k3 + c54*k4, t[i] + c50*h)
k6 = h*f(x[i] + c61*k1 + c62*k2 + c63*k3 + c64*k4 + c65*k5, t[i] + c60*h)
x[i+1] = x[i] + a1*k1 + a3*k3 + a4*k4 + a5*k5
x5 = x[i] + b1*k1 + b3*k3 + b4*k4 + b5*k5 + b6*k6
E = abs(x[n-1]-exact(t[n-1])) #error
print(E)
if E < 10^-5: #error tolerance
return x[n-1] #approximation
print("For dt = 10e-2, x(1) =",Rk4(1.0,0.0,10e-2))
print("For dt = 10e-3, x(1) =",Rk4(1.0,0.0,10e-3))
However, when I run the code, it prints:
5.76914409023e-08
For dt = 10e-2, x(1) = None
4.8151482801e-12
For dt = 10e-3, x(1) = None
Therefore the error, E, is less than 10^-5 in both cases, yet it doesn't print x(1).
^
is bitwise exclusive or in Python, not exponentiation; that's **
. So 10^-5
(10 xor -5) is -15, and none of your errors is less than -15.
You probably want the scientific notation constant 1e-5
or, if you want to write it as an actual exponentiation operation, 10 ** -5
.
For those concerned with performance: the expression using **
is actually compiled to a constant when Python loads the script, since the arguments are both constants, so it won't have any performance impact to write it that way. :-)
The technical post webpages of this site follow the CC BY-SA 4.0 protocol. If you need to reprint, please indicate the site URL or the original address.Any question please contact:yoyou2525@163.com.