Interaction models

This work has been carried out during the 2021 summer as part of an internship at the Centre of Theoretical Physics (CPT) in Marseille under the supervision of Mathieu Génois.

This library implements several models of human face-to-face interactions in a closed environment. The models range from the brownian motion model to some specific pedestrian models.

How to run the code ?

First, you need to choose the model you want to test. You have four choices (They are all described next):

• Brownian motion.
• Ballistic with stop.
• Vicsek model.
• Pedestrian model.

Then you need to import the wanted class. For example, the following code applies if you want to test the Brownian motion:

from brownian import BrownianMotion

Then you need to initialize the class in function of several variables. Many variables are optional. In the case of the Brownian Motion, you need to input two variables: the coefficient of diffusion (diff) and the number of particles (n_particles) in the box. For example, if you want diff = 0.001 and n_particles = 100:

brown = BrownianMotion(0.001, 100)

Then, two options are possible:

• Animate the motion
• Get a tij file

Animation

To animate, you have to import the MovementAnimation class and then launch the animation. There is one variable: the class you want to animate.

from animate import MovementAnimation
   
MovementAnimation(brown)

You should see an animated canvas as shown below:

Tij array

To get the tij array of a model you need to run the following code:

brown_tij = brown.total_movement()

To convert a .dat experimental file into a tij_array you need to run the following code:

import tij

path = 'tij_ICCSS17.dat'
tij_array = tij.conversion(path)

Tij interpreter

The timeline function allows to convert the tij data array into a convenient timeline_array. The interesting data is extracted from the timeline_arrays and is represented in four different graphs (Contact time distribution, Intercontact time distribution, number of contacts dsitribution and weight distribution). One convenient function is the compare_quantities function: it allows to compare your numerical data with experimental data. Here is a code to interpret the data:

pt = 'tij_conf1.dat'
tij_array = tij.conversion(pt)
timeline_array = tij.timeline(tij_array, 20)
quantities_conf = tij.quantities_calculator(timeline_array)

brown = BrownianMotion(0.001, 100)
brown_tij = brown.total_movement()
timeline_array = tij.timeline(brown_tij, 20)
quantities_brown = tij.quantities_calculator(timeline_array)

tij.compare_quantities([quantities_brown, quantities_conf], ['Brownian motion', 'Conf data'],  scale='log')

The output of the code is: