A simulation model is supposed to be a virtual foot print of reality with a lot of assumptions lying underneath. These assumptions can be validated by comparison with physical experiments and data. Once a validated model has been developed and regions of validity have been determined, extrapolations to new parameter ranges can be carried out and performances can be predicted right first time.
Validated simulation models strongly shorten your time to market, eliminate expensive and long duration experiments and prototyping.
Moreover new parameter ranges can be examined as well as that design rules and windows can be derived making the design process efficient and effective.