I work at a corporation and we have engineer's who test "in home" settings which requires a wireless router. Such testing doesn't need to extend beyond the desk of the user.

Currently we keep track of what channels we assign using a wiki and table. However, I realized a few months ago that we never update the table beyond initial entries, so when people move (which is often) the list becomes outdated.

The goal of the table was to avoid saturating any specific channel. So I thought couldn't we just use uniform random numbers to assign channels?

I'm looking at the channels we use, 2.4G and 5.0G. After seeing that certain channels overlap on 2.4 (but not on 5.0G). I thought for 2.4G it would be best to either use the half overlap channels (1, 3, 5, 7, 9, 11)... or the non overlap channels of 1, 6, 11.

We work in multi storied buildings, so there might be bleed over from inbetween floors.

Because of the demand for multiple wireless routers on any given floor (due to testing) results in unavoidable channel overlap. To mitigate any one channel from becoming too overlapped/saturated, I thought random assignment would result in uniform distribution of channels with about equal saturation for each channel.

Thoughts?