PythonModel

A program that simulates crowd dynamics, particularly in a panic situation, using a multi-agent system approach.

Software Used

Python 2.7

Python is designed to be very readable, which is the main reason it was chosen. The previous model was not documented well, so hopefully writing it in Python will help future teams.

We're using Python 2 because there are still some packages that haven't been ported to Python 3. We're not sure which packages we'll be using and don't want to have to rewrite our model if we need a specific package that hasn't been ported yet.

Anaconda

Anaconda is a Python package designed specifically for data analysis. It comes with Spyder, a MATLAB like Python IDE, as well as the most common Python packages for data science.

Download Anaconda here: https://www.continuum.io/downloads

Naming Conventions

Variables - mixedCase

I think this is easily readable and looks better than underscores.

Example:

agentOnePosition = [19,30]

Functions - mixedCase

Typically the function and variables follow the same conventions

Example:

def setAgentPosition(agentNumber):

	agentNumberPosition = [19,30]

Classes - CapWords

Makes it easy to tell that the method (a function that is defined in a class) being called comes from a class

For more https://www.Python.org/dev/peps/pep-0008/

Methodology

Create agents

I think this should be done with an agent class

Assign it x/y coordinates

Preferably it will be in a 2 number list

	Example
	
		agentOnePosition = [5.44,8.32]
		
	I think this is more compact and more functional
	
Could also use two variables 

	Example
	
		agentOneX=5.44
		
		agentOneY=8.32

Process the forces

I'd like to use the Headed Social Force Model (HSFM) because it combines the Social Force Model with the Human Locomotion Model, ideally I'll be able to program all three and we can test them independently

Forces

Agent-agent

Agent-environment

Agent-robot

Agent-goal

	This is what I'm calling the self-driven force

Forces will be lists treated as vectors

Example:

	agent1 acting on agent3 force = [-2.32,5.14]

Agent-agent forces will be equal and opposite so assign the force to the other agent

Example:

	agent3 acting on agent1 force = [2.32,-5.14]

Sum all of the forces for the agent

The final force vector will define their movement direction

QUESTION: Do we just correlate it to their acceleration

Update the agent positions

Move the agents in the based on their final force vector

Check to make sure the agent is within the region

If not then move them as far as possible

CoreySpohn/PythonModel