Added cast function for vectors

This allows to easily change the type of elements in a vector
expression, which can be useful to force a specific type when combining
vectors with different element types. For example, a double-vector can
be cast to float to avoid expensive computations in double precision.