/SIR_modeling

This is a project in which I wrote code to simulation the spread of infection using the SIR model. Analysis of the convergence and stability of the model is done in the report. Additionally, I used a Neuro-Evolution algorithm to design different vaccination models to minimize the amount of infected people while using minimal vaccines

Primary LanguagePython

SIR Modeling

Repository Structure:

  • The report is in the file Report.pdf
  • All the simulation code (including the runner) is in simulation.py
  • parameter_analysis.py contains the code to run the simulation (with no vaccination) over tao within [0,4] and kappa within [1,5] and generate a heatmap of the stopping (convergence times) for analysis.
  • vaccination_polices.py contains some options for the vaccination policies and a way to handle hyper-parameters of vaccination policies.
  • neat_policy.py runs the Neuro Evolution algorithm to create a neural network vaccination policy that will automatically (through Neuro-Evolution) design a policy that maximizes the amount of susceptible people left at the converged solution.
  • vaccination_tester.py handles testing a vaccination policy by running the simulation multiple times with it, finding optimal hyper-parameters, and then plotting the resulting simulation using the optimal vaccination hyper-parameters.
  • init.bat is a windows batch script to create the following directories (required to run the code): results, results/sims, results/analysis, results/models
  • clear_results.bat is a windows batch script to clear all the results in results/sims