Created
October 1, 2023 18:58
-
-
Save Dpananos/55ca8912aeeef12dc7b0fde4181e0b51 to your computer and use it in GitHub Desktop.
Discrete dynamics
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
import matplotlib.pyplot as plt | |
import numpy as np | |
from scipy.stats import multinomial | |
import cmdstanpy | |
def simulate(n_transitions): | |
A = 1000 | |
B = 100 | |
C = 400 | |
N = A + B + C | |
r1 = 0.02 | |
r2 = 0.02 | |
r3 = 0.03 | |
r4 = 0.01 | |
transition_matrix = np.array([ | |
[1-r1, 0, r4], | |
[r1, 1-r2, r3], | |
[0, r2, 1-r4-r3] | |
]) | |
X = np.zeros((n_transitions, 3)) | |
theta = np.zeros((n_transitions, 3)) | |
X[0] = (A, B, C) | |
theta[0] = X[0]/sum(X[0]) | |
for i in range(1, n_transitions): | |
theta[i] = np.linalg.matrix_power(transition_matrix, i) @ theta[0] | |
Aout = multinomial(n=A, p=[1-r1, r1, 0]).rvs().ravel() | |
Bout = multinomial(n=B, p=[0, 1-r2, r2]).rvs().ravel() | |
Cout = multinomial(n=C, p=[r4, r3, 1-r3-r4]).rvs().ravel() | |
A += - np.delete(Aout, 0).sum() + Cout[0] + Bout[0] | |
B += - np.delete(Bout, 1).sum() + Cout[1] + Aout[1] | |
C += - np.delete(Cout, 2).sum() + Bout[2] + Aout[2] | |
X[i] = (A, B, C) | |
return X, theta*N | |
X, t = simulate(365) |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment