Scipy: Speeding up calculation of a 2D complex integral - python

I want to repeatedly calculate a two-dimensional complex integral using dblquad from scipy.integrate. As the number of evaluations will be quite high I would like to increase the evaluation speed of my code.
Dblquad does not seem to be able to handle complex integrands. Thus, I have split the complex integrand into a real and an imaginary part:
def integrand_real(x, y):
R1=sqrt(x**2 + (y-y0)**2 + z**2)
R2=sqrt(x**2 + y**2 + zxp**2)
return real(exp(1j*k*(R1-R2)) * (-1j*z/lam/R2/R1**2) * (1+1j/k/R1))
def integrand_imag(x,y):
R1=sqrt(x**2 + (y-y0)**2 + z**2)
R2=sqrt(x**2 + y**2 + zxp**2)
return imag(exp(1j*k*(R1-R2)) * (-1j*z/lam/R2/R1**2) * (1+1j/k/R1))
y0, z, zxp, k, and lam are variables defind in advance. To evaluate the integral over the area of a circle with radius ra I use the following commands:
from __future__ import division
from scipy.integrate import dblquad
from pylab import *
def ymax(x):
return sqrt(ra**2-x**2)
lam = 0.000532
zxp = 5.
z = 4.94
k = 2*pi/lam
ra = 1.0
res_real = dblquad(integrand_real, -ra, ra, lambda x: -ymax(x), lambda x: ymax(x))
res_imag = dblquad(integrand_imag, -ra, ra, lambda x: -ymax(x), lambda x: ymax(x))
res = res_real[0]+ 1j*res_imag[0]
According to the profiler the two integrands are evaluated about 35000 times. The total calculation takes about one second, which is too long for the application I have in mind.
I am a beginner to scientific computing with Python and Scipy and would be happy about comments that point out ways of improving the evaluation speed. Are there ways of rewriting the commands in the integrand_real and integrand_complex functions that could lead to siginficant speed improvements?
Would it make sense to compile those functions using tools like Cython? If yes: Which tool would best fit this application?

You can gain a factor of about 10 in speed by using Cython, see below:
In [87]: %timeit cythonmodule.doit(lam=lam, y0=y0, zxp=zxp, z=z, k=k, ra=ra)
1 loops, best of 3: 501 ms per loop
In [85]: %timeit doit()
1 loops, best of 3: 4.97 s per loop
This is probably not enough, and the bad news is that this is probably
quite close (maybe factor of 2 at most) to everything-in-C/Fortran speed
--- if using the same algorithm for adaptive integration. (scipy.integrate.quad
itself is already in Fortran.)
To get further, you'd need to consider different ways to do the
integration. This requires some thinking --- can't offer much from
the top of my head now.
Alternatively, you can reduce the tolerance up to which the integral
is evaluated.
# Do in Python
#
# >>> import pyximport; pyximport.install(reload_support=True)
# >>> import cythonmodule
cimport numpy as np
cimport cython
cdef extern from "complex.h":
double complex csqrt(double complex z) nogil
double complex cexp(double complex z) nogil
double creal(double complex z) nogil
double cimag(double complex z) nogil
from libc.math cimport sqrt
from scipy.integrate import dblquad
cdef class Params:
cdef public double lam, y0, k, zxp, z, ra
def __init__(self, lam, y0, k, zxp, z, ra):
self.lam = lam
self.y0 = y0
self.k = k
self.zxp = zxp
self.z = z
self.ra = ra
#cython.cdivision(True)
def integrand_real(double x, double y, Params p):
R1 = sqrt(x**2 + (y-p.y0)**2 + p.z**2)
R2 = sqrt(x**2 + y**2 + p.zxp**2)
return creal(cexp(1j*p.k*(R1-R2)) * (-1j*p.z/p.lam/R2/R1**2) * (1+1j/p.k/R1))
#cython.cdivision(True)
def integrand_imag(double x, double y, Params p):
R1 = sqrt(x**2 + (y-p.y0)**2 + p.z**2)
R2 = sqrt(x**2 + y**2 + p.zxp**2)
return cimag(cexp(1j*p.k*(R1-R2)) * (-1j*p.z/p.lam/R2/R1**2) * (1+1j/p.k/R1))
def ymax(double x, Params p):
return sqrt(p.ra**2 + x**2)
def doit(lam, y0, k, zxp, z, ra):
p = Params(lam=lam, y0=y0, k=k, zxp=zxp, z=z, ra=ra)
rr, err = dblquad(integrand_real, -ra, ra, lambda x: -ymax(x, p), lambda x: ymax(x, p), args=(p,))
ri, err = dblquad(integrand_imag, -ra, ra, lambda x: -ymax(x, p), lambda x: ymax(x, p), args=(p,))
return rr + 1j*ri

Have you considered multiprocessing (multithreading)? It seems that you don't have a need to do a final integration (over the whole set) so simple parallel processing might be the answer. Even if you did have to integrate, you can wait for running threads to finish computation before doing the final integration. That is, you can block the main thread until all workers have completed.
http://docs.python.org/2/library/multiprocessing.html

quadpy (a project of mine) supports many integration schemes for functions over disks. It supports complex-valued functions and is fully vectorized. For example with Peirce's scheme of order 83:
from numpy import sqrt, pi, exp
import quadpy
lam = 0.000532
zxp = 5.0
z = 4.94
k = 2 * pi / lam
ra = 1.0
y0 = 0.0
def f(X):
x, y = X
R1 = sqrt(x ** 2 + (y - y0) ** 2 + z ** 2)
R2 = sqrt(x ** 2 + y ** 2 + zxp ** 2)
return exp(1j * k * (R1 - R2)) * (-1j * z / lam / R2 / R1 ** 2) * (1 + 1j / k / R1)
scheme = quadpy.disk.peirce_1957(20)
val = scheme.integrate(f, [0.0, 0.0], ra)
print(val)
(18.57485726096671+9.619636385589759j)

Related

overflow in exponential function while trying to integrate

I want to numerically integrate a discrete dataset (given ad pandas series) -here orange- which is multiplied with a given analytical exponential function (derivative of a Fermi-Dirac-Distribution) -here blue-. However I fail when the exponent becomes large (e.g. for small T) and thus the derivative fermi_dT(E, mu, T)explodes. I couldn't find a way to rewrite fermi_dT(E, mu, T)in an appropriate way to get it done.
Below is a minimal example (not with pandas series), where I simulated the dataset by a Gaussian.
If T<30. I'll get an overflow. Does anyone see a clever way to get around?
import numpy as np
from scipy import integrate
import matplotlib.pyplot as plt
scale_plot = 1e6
kB = 8.618292134831462e-5 #in eV
Ef = 2.0
def gaussian(E, amp, E0, sig):
return amp * np.exp(-(E-E0)**2 / sig)
def fermi_dT(E, mu, T):
return ((np.exp((E - mu) / (kB * T))*(E-mu)) / ((1 + np.exp((E - mu) / (kB * T)))**2*kB*T**2))
T = 100.0
energies = np.arange(1.,3.,0.001)
plt.plot(energies, (energies-Ef)*fermi_dT(energies, Ef, T))
plt.plot(energies, gaussian(energies, 1e-5, 1.8, .01))
plt.plot(energies, gaussian(energies, 1e-5, 1.8, .01)*(energies-Ef)*fermi_dT(energies, Ef, T)*scale_plot)
plt.show()
cum = integrate.cumtrapz(gaussian(energies, 1e-5, 1.8, .01)*(energies-Ef)*fermi_dT(energies, Ef, T), energies)
print(cum[-1])
This kind of numerical issue is quite usual when dealing with exponential derivatives. The trick is to compute first the log, and only after to apply the exponential:
log(a*exp(b) / (1 + c*exp(d)) ** k) = log(a) + b - k * log(1 + exp(log(c) + d)))
Now, you need to find a way to compute log(1 + exp(x)) accurately. Lucky for you, people have done it before, according to this post. So maybe you could rewrite fermi_dT using log1p:
import numpy as np
def softplus(x, limit=30):
val = np.empty_like(x)
val[x>=limit] = x[x>=limit]
val[x<limit] = np.log1p(np.exp(x[x<limit]))
return val
def fermi_dT(E, mu, T):
a = (E - mu) / (kB * T ** 2)
b = d = (E - mu) / (kB * T)
k = 2
val = np.empty_like(E)
val[E-mu>=0] = np.exp(np.log(a[E-mu>=0]) + b[E-mu>=0] - k * softplus(d[E-mu>=0]))
val[E-mu<0] = -np.exp(np.log(-a[E-mu<0]) + b[E-mu<0] - k * softplus(d[E-mu<0]))
return val

Partial integral in Python

I want to use the Riemann method to evaluate numerically an partial integral in Python. I would like to integrate with respect to x and find a function of t, but i don't know how do this
My fonction : f(x) = cos(2*pi*x*t) its primitive between [-1/2,1/2]: f(t) = sin(pi*t)/t
def riemann(a, b, dx):
if a > b:
a,b = b,a
n = int((b - a) / dx)
s = 0.0
x = a
for i in range(n):
f_i[k] = np.cos(2*np.pi*x)
s += f_i[k]
x += dx
f_i = s * dx
return f_i,t
There's nothing too horrible about your approach. The result does come out close to the true value:
import numpy as np
def riemann(a, b, dx):
if a > b:
a, b = b, a
n = int((b - a) / dx)
s = 0.0
x = a
for i in range(n):
s += np.cos(2 * np.pi * x)
x += dx
return s * dx
print(riemann(0.0, 0.25, 1.0e-3))
print(1 / (2 * np.pi))
0.15965441949277526
0.15915494309189535
Some remarks:
You wouldn't call this Riemann method. It's the midpoint method (of numerical integration).
Pay a little more attention at the boundaries of your domain. Right now, your numerical domain is [a - dx, b + dx].
If you're looking for speed, best collect all your x values (perhaps with linspace), evaluate the function once with all the points, and then np.sum the values up. (Loops in Python are slow.)

What would be the computationally faster way to implement this 2D numerical integration?

I am interested in doing a 2D numerical integration. Right now I am using the scipy.integrate.dblquad but it is very slow. Please see the code below. My need is to evaluate this integral 100s of times with completely different parameters. Hence I want to make the processing as fast and efficient as possible. The code is:
import numpy as np
from scipy import integrate
from scipy.special import erf
from scipy.special import j0
import time
q = np.linspace(0.03, 1.0, 1000)
start = time.time()
def f(q, z, t):
return t * 0.5 * (erf((t - z) / 3) - 1) * j0(q * t) * (1 / (np.sqrt(2 * np.pi) * 2)) * np.exp(
-0.5 * ((z - 40) / 2) ** 2)
y = np.empty([len(q)])
for n in range(len(q)):
y[n] = integrate.dblquad(lambda t, z: f(q[n], z, t), 0, 50, lambda z: 10, lambda z: 60)[0]
end = time.time()
print(end - start)
Time taken is
212.96751403808594
This is too much. Please suggest a better way to achieve what I want to do. I tried to do some search before coming here, but didn't find any solution. I have read quadpy can do this job better and very faster but I have no idea how to implement the same. Please help.
You could use Numba or a low-level-callable
Almost your example
I simply pass function directly to scipy.integrate.dblquad instead of your method using lambdas to generate functions.
import numpy as np
from scipy import integrate
from scipy.special import erf
from scipy.special import j0
import time
q = np.linspace(0.03, 1.0, 1000)
start = time.time()
def f(t, z, q):
return t * 0.5 * (erf((t - z) / 3) - 1) * j0(q * t) * (1 / (np.sqrt(2 * np.pi) * 2)) * np.exp(
-0.5 * ((z - 40) / 2) ** 2)
def lower_inner(z):
return 10.
def upper_inner(z):
return 60.
y = np.empty(len(q))
for n in range(len(q)):
y[n] = integrate.dblquad(f, 0, 50, lower_inner, upper_inner,args=(q[n],))[0]
end = time.time()
print(end - start)
#143.73969149589539
This is already a tiny bit faster (143 vs. 151s) but the only use is to have a simple example to optimize.
Simply compiling the functions using Numba
To get this to run you need additionally Numba and numba-scipy. The purpose of numba-scipy is to provide wrapped functions from scipy.special.
import numpy as np
from scipy import integrate
from scipy.special import erf
from scipy.special import j0
import time
import numba as nb
q = np.linspace(0.03, 1.0, 1000)
start = time.time()
#error_model="numpy" -> Don't check for division by zero
#nb.njit(error_model="numpy",fastmath=True)
def f(t, z, q):
return t * 0.5 * (erf((t - z) / 3) - 1) * j0(q * t) * (1 / (np.sqrt(2 * np.pi) * 2)) * np.exp(
-0.5 * ((z - 40) / 2) ** 2)
def lower_inner(z):
return 10.
def upper_inner(z):
return 60.
y = np.empty(len(q))
for n in range(len(q)):
y[n] = integrate.dblquad(f, 0, 50, lower_inner, upper_inner,args=(q[n],))[0]
end = time.time()
print(end - start)
#8.636585235595703
Using a low level callable
The scipy.integrate functions also provide the possibility to pass C-callback function instead of a Python function. These functions can be written for example in C, Cython or Numba, which I use in this example. The main advantage is, that no Python interpreter interaction is necessary on function call.
An excellent answer of #Jacques Gaudin shows an easy way to do this including additional arguments.
import numpy as np
from scipy import integrate
from scipy.special import erf
from scipy.special import j0
import time
import numba as nb
from numba import cfunc
from numba.types import intc, CPointer, float64
from scipy import LowLevelCallable
q = np.linspace(0.03, 1.0, 1000)
start = time.time()
def jit_integrand_function(integrand_function):
jitted_function = nb.njit(integrand_function, nopython=True)
#error_model="numpy" -> Don't check for division by zero
#cfunc(float64(intc, CPointer(float64)),error_model="numpy",fastmath=True)
def wrapped(n, xx):
ar = nb.carray(xx, n)
return jitted_function(ar[0], ar[1], ar[2])
return LowLevelCallable(wrapped.ctypes)
#jit_integrand_function
def f(t, z, q):
return t * 0.5 * (erf((t - z) / 3) - 1) * j0(q * t) * (1 / (np.sqrt(2 * np.pi) * 2)) * np.exp(
-0.5 * ((z - 40) / 2) ** 2)
def lower_inner(z):
return 10.
def upper_inner(z):
return 60.
y = np.empty(len(q))
for n in range(len(q)):
y[n] = integrate.dblquad(f, 0, 50, lower_inner, upper_inner,args=(q[n],))[0]
end = time.time()
print(end - start)
#3.2645838260650635
Generally it is much, much faster to do a summation via matrix operations than to use scipy.integrate.quad (or dblquad). You could rewrite your f(q, z, t) to take in a q, z and t vector and return a 3D-array of f-values using np.tensordot, then multiply your area element (dtdz) with the function values and sum them using np.sum. If your area element is not constant, you have to make an array of area-elements and use np.einsum To take your integration limits into account you can use a masked array to mask the function values outside your integration limits before summarizing. Take note that np.einsum overlooks the masks, so if you use einsum you can use np.where to set function values outside your integration limits to zero. Example (with constant area element and simple integration limits):
import numpy as np
import scipy.special as ss
import time
def f(q, t, z):
# Making 3D arrays before computation for readability. You can save some time by
# Using tensordot directly when computing the output
Mq = np.tensordot(q, np.ones((len(t), len(z))), axes=0)
Mt = np.tensordot(np.ones(len(q)), np.tensordot(t, np.ones(len(z)), axes = 0), axes = 0)
Mz = np.tensordot(np.ones((len(q), len(t))), z, axes = 0)
return Mt * 0.5 * (ss.erf((Mt - Mz) / 3) - 1) * (Mq * Mt) * (1 / (np.sqrt(2 * np.pi) * 2)) * np.exp(
-0.5 * ((Mz - 40) / 2) ** 2)
q = np.linspace(0.03, 1, 1000)
t = np.linspace(0, 50, 250)
z = np.linspace(10, 60, 250)
#if you have constand dA you can shave some time by computing dA without using np.diff
#if dA is variable, you have to make an array of dA values and np.einsum instead of np.sum
t0 = time.process_time()
dA = np.diff(t)[0] * np.diff(z)[0]
func_vals = f(q, t, z)
I = np.sum(func_vals * dA, axis=(1, 2))
t1 = time.process_time()
this took 18.5s on my 2012 macbook pro (2.5GHz i5) with dA = 0.04. Doing things this way also allows you to easily choose between precision and efficiency, and to set dA to a value that makes sense when you know how your function behaves.
However, it is worth noting that if you want a larger amount of points, you have to split up your integral, or else you risk maxing out your memory (1000 x 1000 x 1000) doubles requires 8GB of ram. So if you are doing very big integrations with high presicion it can be worth doing a quick check on the memory required before running.

Solving a boundary value problem DE in python

I am trying to solve the following set of DE's:
dx' = cos(a)
dy' = sin(a)
dF' = - b * x * cos(a) + sin(a)
da' = (b * x * sin(a) + cos(a)) / F
with the conditions:
x(0) = y(0) = x(1) = 0
y(1) = 0.6
F(0) = 0.38
a(0) = -0.5
I tried following a similar problem, but I just can't get it to work. Is it possible, that my F(0) and a(0) are completely off, I am not even sure about them.
import numpy as np
from scipy.integrate import solve_bvp
import matplotlib.pyplot as plt
beta = 5
def fun(x, y):
x, dx, y, dy, F, dF, a, da, = y;
dxds=np.cos(a)
dyds=np.sin(a)
dFds=-beta * x * np.cos(a) + np.sin(a)
dads=(beta * x * np.sin(a) + np.cos(a) ) / F
return dx, dxds, dy, dyds, dF, dFds, da, dads
def bc(ya, yb):
return ya[0], yb[0], ya[2], yb[2] + 0.6, ya[4] + 1, yb[4] + 1, ya[6], yb[6]
x = np.linspace(0, 0.5, 10)
y = np.zeros((8, x.size))
y[4] = 0.38
y[6] = 2.5
res = solve_bvp(fun, bc, x, y)
print(res.message)
x_plot = np.linspace(0, 0.5, 200)
plt.plot(x_plot, res.sol(x_plot)[0])
I think that you have foremost a physics problem, translating the physical situation into an ODE system.
x(s) and y(s) are the coordinates of the rope where s is the length along the rope. Consequently, (x'(s),y'(s)) is a unit vector that is uniquely characterized by its angle a(s), giving
x'(s) = cos(a(s))
y'(s) = sin(a(s))
To get the shape, one now has to consider the mechanics. The assumption seems to be that the rope rotates without spiraling around the rotation axis, staying in one plane. Additionally, from the equilibrium of forces you also get that the other two equations are indeed first order, not second order equations. So your state only has 4 components and the ODE system function thus has to be
def fun(s, u):
x, y, F, a = u;
dxds=np.cos(a)
dyds=np.sin(a)
dFds=-beta * x * np.cos(a) + np.sin(a)
dads=(beta * x * np.sin(a) + np.cos(a) ) / F
return dxds, dyds, dFds, dads
Now there are only 4 boundary condition slots available, which are the coordinates of the start and end of the rope.
def bc(ua, ub):
return ua[0], ub[0], ua[1], ub[1] - 0.6
Additionally, the interval length for s is also the rope length, so a value of 0.5 is impossible for the given coordinates on the pole, try 1.0. There is some experimentation needed to get an initial guess that does not lead to a singular Jacobian in the BVP solver. In the end I get the solution in the x-y plane
with the components

Numpy/Scipy Solve simulataneous equations with integrals in them

I am trying to use numpy and scipy to solve the following two equations:
P(z) = sgn(-cos(np.pi*D1) + cos(5*z)) * sgn(-cos(np.pi*D2) + cos(6*z))
1. 0 = 2/2pi ∫ P(z,D1,D2) * cos(5z) dz + z/L
2. 0 = 2/2pi ∫ P(z,D1,D2) * cos(6z) dz - z/L
for D1 and D2 (integral limits are 0 -> 2pi).
My code is:
def equations(p, z):
D1, D2 = p
period = 2*np.pi
P1 = lambda zz, D1, D2: \
np.sign(-np.cos(np.pi*D1) + np.cos(6.*zz)) * \
np.sign(-np.cos(np.pi*D2) + np.cos(5.*zz)) * \
np.cos(6.*zz)
P2 = lambda zz, D1, D2: \
np.sign(-np.cos(np.pi*D1) + np.cos(6.*zz)) * \
np.sign(-np.cos(np.pi*D2) + np.cos(5.*zz)) * \
np.cos(5.*zz)
eq1 = 2./period * integrate.quad(P1, 0., period, args=(D1,D2), epsabs=0.01)[0] + z
eq2 = 2./period * integrate.quad(P2, 0., period, args=(D1,D2), epsabs=0.01)[0] - z
return (eq1, eq2)
z = np.arange(0., 1000., 0.01)
N = int(len(z))
D1 = np.empty([N])
D2 = np.empty([N])
for i in range(N):
D1[i], D2[i] = fsolve(equations, x0=(0.5, 0.5), args=z[i])
print D1, D2
Unfortunately, it does not seem to converge. I don't know much about numerical methods and was hoping someone could give me a hand.
Thank you.
P.S. I'm also trying the following which should be equivalent:
import numpy as np
from scipy.optimize import fsolve
from scipy import integrate
from scipy import signal
def equations(p, z):
D1, D2 = p
period = 2.*np.pi
K12 = 1./L * z
K32 = -1./L * z + 1.
P1 = lambda zz, D1, D2: \
signal.square(6.*zz, duty=D1) * \
signal.square(5.*zz, duty=D2) * \
np.cos(6.*zz)
P2 = lambda zz, D1, D2: \
signal.square(6.*zz, duty=D1) * \
signal.square(5.*zz, duty=D2) * \
np.cos(5.*zz)
eq1 = 2./period * integrate.quad(P1, 0., period, args=(D1,D2))[0] + K12
eq2 = 2./period * integrate.quad(P2, 0., period, args=(D1,D2))[0] - K32
return (eq1, eq2)
h = 0.01
L = 10.
z = np.arange(0., L, h)
N = int(len(z))
D1 = np.empty([N])
D2 = np.empty([N])
for i in range(N):
D1[i], D2[i] = fsolve(equations, x0=(0.5, 0.5), args=z[i])
print
print z[i]
print ("%0.8f,%0.8f" % (D1[i], D2[i]))
print
PSS:
I implemented what you wrote (I think I understand it!), very nicely done. Thank you. Unfortunately, I really don't have much skill in this field and don't really know how to make a suitable guess, so I just guess 0.5 (I also added a small amount of noise to the initial guess to try and improve it). The result I'm getting have numerical errors it seems, and I'm not sure why, I was hoping you could point me in the right direction. So essentially, I did an FFT sweep (did an FFT for each dutycycle variation and looked at the frequency component at 5, which is shown below in the graph) and found that the linear part (z/L) is slightly jagged.
PSSS:
Thank you for that, I've noted some of the techniques you've suggested. I tried replicated your second graph as it seems very useful. To do this, I kept D1 (D2) fixed and swept D2 (D1), and I did this for various z values. fmin did not always find the correct minimum (it was dependent on the initial guess) so I swept the initial guess of fmin until I found the correct answer. I get a similar answer to you. (I think it's correct?)
Also, I would just like to say that you might like to give me your contact details, as this solution as a step in finding the solution to a problem I have (I'm a student doing research), and I will most certainly acknowledge you in any papers in which this code is used.
#!/usr/bin/env python
import numpy as np
from scipy.optimize import fsolve
from scipy import integrate
from scipy import optimize
from scipy import signal
######################################################
######################################################
altsigns = np.ones(50)
altsigns[1::2] = -1
def get_breaks(x, y, a, b):
sa = np.arange(0, 2*a, 2)
sb = np.arange(0, 2*b, 2)
zx = (( x + sa) % (2*a))*np.pi/a
zx2 = ((-x + sa) % (2*a))*np.pi/a
zy = (( y + sb) % (2*b))*np.pi/b
zy2 = ((-y + sb) % (2*b))*np.pi/b
zi = np.r_[np.sort(np.hstack((zx, zx2, zy, zy2))), 2*np.pi]
if zi[0]:
zi = np.r_[0, zi]
return zi
def integrals(x, y, a, b):
zi = get_breaks(x % 1., y % 1., a, b)
sins = np.vstack((np.sin(b*zi), np.sin(a*zi)))
return (altsigns[:zi.size-1]*(sins[:,1:] - sins[:,:-1])).sum(1) / np.array((b, a))
def equation1(p, z, d2):
D2 = d2
D1 = p
I1, _ = integrals(D1, D2, deltaK1, deltaK2)
eq1 = 1. / np.pi * I1 + z
return abs(eq1)
def equation2(p, z, d1):
D1 = d1
D2 = p
_, I2 = integrals(D1, D2, deltaK1, deltaK2)
eq2 = 1. / np.pi * I2 - z + 1
return abs(eq2)
######################################################
######################################################
z = [0.2, 0.4, 0.6, 0.8, 1.0]#np.arange(0., 1., 0.1)
step = 0.05
deltaK1 = 5.
deltaK2 = 6.
f = open('data.dat', 'w')
D = np.arange(0.0, 1.0, step)
D1eq1 = np.empty([len(D)])
D2eq2 = np.empty([len(D)])
D1eq1Err = np.empty([len(D)])
D2eq2Err = np.empty([len(D)])
for n in z:
for i in range(len(D)):
# Fix D2 and solve for D1.
for guessD1 in np.arange(0.,1.,0.1):
D2 = D
tempD1 = optimize.fmin(equation1, guessD1, args=(n, D2[i]), disp=False, xtol=1e-8, ftol=1e-8, full_output=True)
if tempD1[1] < 1.e-6:
D1eq1Err[i] = tempD1[1]
D1eq1[i] = tempD1[0][0]
break
else:
D1eq1Err[i] = -1.
D1eq1[i] = -1.
# Fix D1 and solve for D2.
for guessD2 in np.arange(0.,1.,0.1):
D1 = D
tempD2 = optimize.fmin(equation2, guessD2, args=(n, D1[i]), disp=False, xtol=1e-8, ftol=1e-8, full_output=True)
if tempD2[1] < 1.e-6:
D2eq2Err[i] = tempD2[1]
D2eq2[i] = tempD2[0][0]
break
else:
D2eq2Err[i] = -2.
D2eq2[i] = -2.
for i in range(len(D)):
f.write('%0.8f,%0.8f,%0.8f,%0.8f,%0.8f\n' %(D[i], D1eq1[i], D2eq2[i], D1eq1Err[i], D2eq2Err[i]))
f.write('\n\n')
f.close()
This is a very ill-posed problem. Let's recap what you are trying to do:
You want to solve 100000 optimization problems
Each optimization problem is 2 dimensional, so you need O(10000) function evaluations (estimating O(100) function evaluations for a 1D optimization problem)
Each function evaluation depends on the evaluation of two numerical integrals
The integrands contain jumps, i.e. they are 0-times contiguously differentiable
The integrands are composed of periodic functions, so they have multiple minima and maxima
So you are off to a very hard time. In addition, even in the most optimistic estimate in which all factors in the integrand that are < 1 are replaced by 1, the integrals can only take values between -2*pi and 2*pi. Much less than that in reality. So you can already see that you only have a chance of a solution for
I1 - z = 0
I2 + z = 0
for very small numbers of z. So there is no point in trying up to z = 1000.
I am almost certain that this is not the problem you need to solve. (I cannot imagine a context in which such a problem would appear. It seems like a weird twist on Fourier coefficient computation...) But in case you insist, your best bet is to work on the inner loop first.
As you noted, the numerical evaluation of the integrals is subject to large errors. This is due to the jumps introduced by the sgn() function. Functions such as scipy.integrate.quad() tend to use higher order algorithms which assume that the integrands are smooth. If they are not, they perform very badly. You either need to hand-pick an algorithm that can deal with jumps or, much better in this case, do the integrals by hand:
The following algorithm calculates the jump points of the sgn() function and then evaluates the analytic integrals on all pieces:
altsigns = np.ones(50)
altsigns[1::2] = -1
def get_breaks(x, y, a, b):
sa = np.arange(0, 2*a, 2)
sb = np.arange(0, 2*b, 2)
zx = (( x + sa) % (2*a))*np.pi/a
zx2 = ((-x + sa) % (2*a))*np.pi/a
zy = (( y + sb) % (2*b))*np.pi/b
zy2 = ((-y + sb) % (2*b))*np.pi/b
zi = np.r_[np.sort(np.hstack((zx, zx2, zy, zy2))), 2*pi]
if zi[0]:
zi = np.r_[0, zi]
return zi
def integrals(x, y, a, b):
zi = get_breaks(x % 1., y % 1., a, b)
sins = np.vstack((np.sin(b*zi), np.sin(a*zi)))
return (altsigns[:zi.size-1]*(sins[:,1:] - sins[:,:-1])).sum(1) / np.array((b, a))
This gets rid of the problem of the numerical integration. It is very accurate and fast. However, even the integrals will not be perfectly contiguous for all parameters, so in order to solve your optimization problem, you are better off using an algorithm that doesn't rely on the existence of any derivatives. The only choice in scipy is scipy.optimize.fmin(), which you can use like:
def equations2(p, z):
x, y = p
I1, I2 = integrals(x, y, 6., 5.)
fact = 1. / pi
eq1 = fact * I1 + z
eq2 = fact * I2 - z
return eq1, eq2
def norm2(p, z):
eq1, eq2 = equations2(p, z)
return eq1**2 + eq2**2 # this has the minimum when eq1 == eq2 == 0
z = 0.25
res = fmin(norm2, (0.25, 0.25), args=(z,), xtol=1e-8, ftol=1e-8)
print res
# -> [ 0.3972 0.5988]
print equations2(res, z)
# -> (-2.7285737558280232e-09, -2.4748670890417657e-09)
You are still left with the problem of finding suitable starting values for all z, which is still a tricky business. Good Luck!
Edit
To check if you still have numerical errors, plug the result of the optimization back in the equations and see if they are satisfied to the required accuracy, which is what I did above. Note that I used (0.25, 0.25) as a starting value, since starting at (0.5, 0.5) didn't lead to convergence. This is normal for optimizations problems with local minima (such as yours). There is no better way to deal with this other than trying multiple starting values, rejecting non-converged results. In the case above, if equations2(res, z) returns anything higher than, say, (1e-6, 1e-6), I would reject the result and try again with a different starting value. A very useful technique for successive optimization problems is to use the result of the previous problem as the starting value for the next problem.
Note however that you have no guarantee of a smooth solution for D1(z) and D2(z). Just a tiny change in D1 could push one break point off the integration interval, resulting in a big change of the value of the integral. The algorithm may well adjust by using D2, leading to jumps in D1(z) and D2(z). Note also that you can take any result modulo 1, due to the symmetries of cos(pi*D1).
The bottom line: There shouldn't be any remaining numerical inaccuracies if you use the analytical formula for the integrals. If the residuals are less than the accuracy you specified, this is your solution. If they are not, you need to find better starting values. If you can't, a solution may not exist. If the solutions are not contiguous as a function of z, that is also expected, since your integrals are not contiguous. Good luck!
Edit 2
It appears your equations have two solutions in the interval z in [0, ~0.46], and no solutions for z > 0.46, see the first figure below. To prove this, see the good old graphical solution in the second figure below. The contours represent solutions of Eq. 1 (vertical) and Eq. 2 (horizontal), for different z. You can see that the contours cross twice for z < 0.46 (two solutions) and not at all for z > 0.46 (no solution that simultaneously satisfies both equations). If this is not what you expected, you need to write down different equations (which was my suspicion in the first place...)
Here is the final code I was using:
import numpy as np
from numpy import sin, cos, sign, pi, arange, sort, concatenate
from scipy.optimize import fmin
a = 6.0
b = 5.0
def P(z, x, y):
return sign((cos(a*z) - cos(pi*x)) * (cos(b*z) - cos(pi*y)))
def P1(z, x, y):
return P(z, x, y) * cos(b*z)
def P2(z, x, y):
return P(z, x, y) * cos(a*z)
altsigns = np.ones(50)
altsigns[1::2] = -1
twopi = 2*pi
pi_a = pi/a
da = 2*pi_a
pi_b = pi/b
db = 2*pi_b
lim = np.array([0., twopi])
def get_breaks(x, y):
zx = arange(x*pi_a, twopi, da)
zx2 = arange((2-x)*pi_a, twopi, da)
zy = arange(y*pi_b, twopi, db)
zy2 = arange((2-y)*pi_b, twopi, db)
zi = sort(concatenate((lim, zx, zx2, zy, zy2)))
return zi
ba = np.array((b, a))[:,None]
fact = np.array((1. / b, 1. / a))
def integrals(x, y):
zi = get_breaks(x % 1., y % 1.)
sins = sin(ba*zi)
return fact * (altsigns[:zi.size-1]*(sins[:,1:] - sins[:,:-1])).sum(1)
def equations2(p, z):
x, y = p
I1, I2 = integrals(x, y)
fact = 1. / pi
eq1 = fact * I1 + z
eq2 = fact * I2 - z
return eq1, eq2
def norm2(p, z):
eq1, eq2 = equations2(p, z)
return eq1**2 + eq2**2
def eval_integrals(Nx=100, Ny=101):
x = np.arange(Nx) / float(Nx)
y = np.arange(Ny) / float(Ny)
I = np.zeros((Nx, Ny, 2))
for i in xrange(Nx):
xi = x[i]
Ii = I[i]
for j in xrange(Ny):
Ii[j] = integrals(xi, y[j])
return x, y, I
def solve(z, start=(0.25, 0.25)):
N = len(z)
res = np.zeros((N, 2))
res.fill(np.nan)
for i in xrange(N):
if i < 100:
prev = start
prev = fmin(norm2, prev, args=(z[i],), xtol=1e-8, ftol=1e-8)
if norm2(prev, z[i]) < 1e-7:
res[i] = prev
else:
break
return res
#x, y, I = eval_integrals(Nx=1000, Ny=1001)
#zlvl = np.arange(0.2, 1.2, 0.2)
#contour(x, y, -I[:,:,0].T/pi, zlvl)
#contour(x, y, I[:,:,1].T/pi, zlvl)
N = 1000
z = np.linspace(0., 1., N)
res = np.zeros((N, 2, 2))
res[:,0,:] = solve(z, (0.25, 0.25))
res[:,1,:] = solve(z, (0.05, 0.95))

Categories

Resources