Sure, but the best answer really depends on what you're trying to do. If you're looking to pass an arbitrary array of thresholds, then you could just construct the functor over a reference to a vector or array.

If you're trying to associate each edge? vertex? with a threshold, then you could do one of two things. You could basically take the direct approach (as above) and construct the functor over a reference to a vector or unordered_map (depending on the properties of your graph). Alternatively, you could construct the functor over a /property map/, which abstracts the read/write operations to a container (vector or hash table) defined elsewhere in your program. See the BGL documentation about exterior properties. It's light, but it'll get you started in the right direction.

Either of the two cases is appropriate. If you plan on reusing your functor in a generic context, I would use the second approach.