I know that bayesian optimisation is a strategy for optimising black-box functions. But if i have some information about type of function or it's specific behaviour at some interval what methods of accounting this information do i have?

As I understand i need to specify it in surrogate function. I use gaussian processes.