This is an old final project I did with Sarah Howerter from a Modeling Complex Systems class at UVM that ran during the Fall of 2018. Isolated from (the old repo) because it's still cool to me. We created a nonspatial, instantaneous model for voting under various election systems in a world where voters followed their gut and voted based on how agreeable their 'opinion vector' was to that of the candidates. Using our model, we wanted to investigate how effective each voting system is at maximizing overall happiness of a population.
- Checkout the report here.
- Things I would change at a glance:
- Agreeability metric should be something more traditional like dot product or a norm.
- Opinion space should be something more interesting than a multidimensional uniform blob, like a multinomial Gaussian.
- Candidates could be more complex than just randomly sampled people