I26 vaccination age struct

[I-Bouros]

Issue

• [x] Which issue does your pull request address? If none of the existing issues are applicable, you can write a new issue.

Testing, documentation, and checks

• [x] If your pull request introduces new code, you should write automated tests (explained here).
• [x] Check that all tests pass using devtools::check().
• [x] Check compliance with the preferred R style using lintr::lint_package().
• [x] Build documentation using devtools::document().
• [x] If you changed the readme, run knitr::knit("README.Rmd").

Pull request

• [x] Assign at least one reviewer.

[I-Bouros]

If it's not too difficult to figure out reasonable values of nu and coverage for this---Could you add to the notebook a pair of simulations comparing two interventions, both with approximately the same total rate of vaccination, but in one case applied equally to all age groups, and in the other applied to each age group in succession starting with the oldest. This would make most sense with an age structured death rate similar to the contact matrix notebook https://github.com/Como-DTC-Collaboration/como-models-math-biosci/blob/e1ffcc491eaec2a7d6ce1b5a0d99ee548f1a49af/contact_matrix/SEIRD_contact_matrix.Rmd#L194

You would show that the second simulation has fewer total deaths, even though the population vaccination rates are comparable. This is not a surprising result, but it is a highly relevant one, and justifies the complexity of the model.

I had a discussion with Ben yesterday and while he agrees with the idea of having a time-dependent nu parameter as we have discussed before, it might be better to postpone that idea for a bit.