couzinswarm

Simulate fish swarming behavior with the model by Iain Couzin et al. Use with caution as I can't seem to replicate some behavior (e.g. the torus swarming).

Example

from couzinswarm import Swarm

swarm = Swarm()
r, v = swarm.simulate(1000)

Install

pip install couzinswarm

Elaborate example

import numpy as np
import matplotlib.pyplot as pl
from mpl_toolkits.mplot3d import Axes3D
from couzinswarm import Swarm

# note: the dimension of space is measure in fish length,
# such that r = 1 means a length of one fish

swarm = Swarm(
             number_of_fish=20,
             repulsion_radius=1,
             orientation_width=10,
             attraction_width=10,
             # this angle is given in radians
             # up to np.pi (not 360 degrees as in
             # the paper)
             angle_of_perception=np.pi,
             # radians per unit of time
             turning_rate=0.1,
             # fish lengths per unit of time
             speed=0.1,
             # in units of fish length
             noise_sigma=0.1,
             dt=0.1,
             # geometry of box
             box_lengths=[100,100,100],
             # boundary conditions
             reflect_at_boundary = [True, True, True],
             verbose=False,
             )



fig = pl.figure()
ax = fig.add_subplot(111, projection='3d')

N_t = 1000

t = np.arange(N_t+1)

# Note that r.shape = v.shape = ( N_fish, N_t+1, 3 )
positions, directions = swarm.simulate(N_t)
r, v = positions, directions

for i in range(swarm.number_of_fish):
    ax.plot(r[i,:,0], r[i,:,1], r[i,:,2])

pl.show()