# TVizir2 and TVizirIsland

The `TVizir2` and `TVizirIsland` classes are the `TMaster` subclass for global optimizer, the former 
being the more commonly used disregarding the chosen solver and runner. The latter is dedicated to 
the *island* principle: instead of growing a large population, $N_{\rm Is}$ populations are defined 
(this parameter can be changed by the dedicated `setIsland(int)` method) and are growing 
independently, exchanging inhabitants from time to time.

The `setSize` may be used to configure this. The first argument defines the population size (default 
to 250). The second optional argument defines the maximum number of evaluations that the solver may 
use (default to 100 000). The third optional argument defines the number of items born to create a 
new generation (default to 50).