In the study of biomolecules as well as in many other fields of science, computer simulations are pervasively used to solve difficult problems. However, very often the systems complexity makes the application of computer simulations challenging. Many systems of interest exhibit long lived metastable states separated by kinetic bottlenecks. In such cases, only very rarely occurring fluctuations allow the system to leave a metastable state. This makes the transitions from one state to another rare events. However, although rare, these events are crucial for a correct description of the system. For instance, phenomena such as nucleation, chemical reactions, and protein folding are a few examples of rare events. Unfortunately, the time scale of standard simulation falls short of what needed and the simulation of rare events is one of the main challenges of present day simulations. Here we present a novel approach to this problem, based on the introduction of a variational principle. We show how this variational principle can be used to study complex problems and calculate transition rates of rare events. We underline that besides offering computational efficiency this new approach provides a qualitative new point of view that will have far reaching consequences in the future.