Re: voltage drop
There is another aspect to this problem that is worth some attention. I just did a quick calculation., using an Excel spreadsheet, using the numbers you gave. It told me the voltage drop was 22%.
However, that presumed that the entire 15 amps of load was located at the end of the 450 foot run, and that is simply not true. You could calculate the VD from the source to the first light at the full 15 amps, and then calculate the VD from the first light to the second light at a slightly lower current, and continue the process until you reach the last light in the string, which will probably have a load of less than one amp. But this is a very messy process.
So I would start by determining how much of the total 450 foot run is traversed by all of the 15 amps. That is, does all 15 amps travel through the 150 foot distance between the doors? Does all 15 amps travel through some additional distance from the source to the first switch? Pick a number that represents the minimum distance from the source to the first light (i.e., after which you can say that the load further downstream is lower than 15 amps). Do a calculation for that distance, using a 15 amp load.
As an example, with a distance of 150 feet and a load of 15 amps, and presuming a single 120 volt circuit, I calculated that you need a minimum of #8 AWG, in order to limit the voltage drop to 3%. With the same distance and current, but with a full boat (i.e., a 3-phase, 208 volt run), I calculated that you need a minimum of #10 AWG, in order to limit the voltage drop to 3%.