I need advice on a possible approach that I can use to reduce the number of simulations that I need to run.

At the moment I have a simulation parameter (number of a certain device, x) and I need to evaluate how my simulation will perform with a varied number of these devices e.g. 2, 4, 6 ,8 10. However if I choose the 4 device scenario, the 4 devices can be in many different positions (495 to be precise) and depending on the positions the 4 devices are in, I may get slightly different results.

If I run all the possibilities for all the varying number of devices, it will be take me absolutely ages to get my results.

Is there an approach that I can use to reduce the number of simulation runs while maintaining the statistical validity of my results? I've been reading about fractional factorial analysis but this usually seems to have multiple factors with a few levels per factor. From what I can see of my problem I have one factor with a huge number of levels.

Any help appreciated.