fix immunized

[GianmarcoRasi]

Multiply 'fraction' by the proportion of susceptible individual to be sure to subtract and add a value that does not exceeds the proportion of susceptible individials

[AllenDowney]

I see the issue you are addressing -- this function is error-prone if fraction exceeds s. But the intention is that fraction is the fraction of the population that is immunized, so the responsibility is on the caller to make sure the argument is sensible. I'm inclined to leave it as is. What do you think?

[AllenDowney]

BTW, ignore the failing test -- that has nothing to do with this PR.

[GianmarcoRasi]

Thank you for answering, and for asking about my opinion.

I think that sweeping over fraction would result in incorrect estimates of the count of overall infected individuals, since when s < fraction <= 1 the amount of susceptible individuals becomes negative. So, if we want to leave the function as it is then we should sweep between $0$ and the fraction of susceptible individuals s, not between $0$ and $1$.

However, leaving the code untouched would make it easier for a reader to internalize the concept (as it worked for me) even if the result is not 100% correct.