Initialisation from more advanced state

Currently simulations are started from randomly selected three infected persons. In particular, in v3 we have much more detail, and it would be possible to initialise simulations from known region- and sector-specific prevalence statistics.