Retired winter 2024. Too messy, not a realistic application of subclassing. We were using subclassing and inheritance in contexts in which I would use configuration files and possibly functional programming techniques in a "for real" application. An SIR model might still be a good exercise, but not this SIR model at this point in our CS1/CS2 series.
A simple grid model of contagion
Rumors, diseases, and memes have somewhat different modes of transmission, but follow similar patterns of spread from person to person. Models (simulations) can help us consider how the spread and how their spread may be altered.
Compartmental models are a basic tool in epidemiology. SIR ( susceptible-infected-recovered) models are among the simplest compartmental models. They model potential transitions of a set of individuals or homogenous regions between states of susceptibility (S), infection (I), and recovery (R).
In an SIR model, a susceptible individual may become an infected individual through contact with another infected individual. An infected individual may recover. Recovered individuals gain immunity and are therefore not susceptible to reinfection. An infected person might instead die, in which case they are also not susceptible (and so death is technically a subcategory of recovery).
SIR models can be conveniently modeled by a main loop in which a simulated clock or calendar moves forward in regular increments. These regular increments, or "ticks" of the virtual clock, provide an approximation of continuous time in which many individuals may be evolving concurrently. To maintain an illusion that all individuals evolve simultaneously, each simulation cycle (tick) is divided into two parts. In the first part, a next state for each individual is computed, using only the current states of other individuals. (Thus the order in which we loop through individuals does not matter.) In the second part, each individual advances to its next state (and again the order in which we loop through individuals doesn't matter).
While our contagion model is designed to illustrate techniques of computer simulation, and particularly SIR models of contagion, it is not based on a real model from epidemiology. The numerical parameters have been selected merely to provide interesting, somewhat realistic behavior, and do not reflect actual numbers (e.g., infection rates, incubation periods, recovery times and rates) for any particular disease. It is an illustration, not a real epidemiological model. Nonetheless I think it may give you some insight into how some kinds of simulation model are constructed.
Detailed instructions are in doc/HOWTO.md. After constructing the
simulation with two classes of individual Typical and AtRisk, you
should construct a new class of individual Wanderer
and incorporate it in the simulation. This will require changes in
several places, such as initializing the
Population object with a suitable proportion of Wanderer individuals
based on numerical parameters in contagion.ini. The amount of code is
small, and no individual piece of code is complex, but you must
understand how all the pieces fit together to succeed in adding a piece.
Turn in three files.
model.py, a Python source file. This file must contain a class calledWandererso that the person grading the project can easily find it!contagion.ini. This configuration file should contain a section[Wanderer]that sets parameters for your new classcontagion.png, a screen-shot of the display produced by your model running with the provided configuration file.