Ronboco,
My professional background was in chemical process research and development. I learned that it was utterly futile to begin experimental work until I had a reliable analytical technique that would inform me of the result of the experiment. With respect to audio, we all use our ears, and our ears should always remain the final arbiter. The problem with listening is that we all suffer from fatigue, at which point we are not listening critically. We also tend to focus on a single variable rather than making a broad based assessment when trying to optimize by ear. On the other hand, I can work on my room for hours at a time using REW to assess the impact of a given change on overall frequency response, ringing, the balance of direct to reflected sound, early reflection mitigation, etc. I can go back a month later and look at the data and draw the same conclusions, or even discern something I missed the first time. So yes, measure first, attempt to fix a given problem optimizing existing gear second, and spend money third. However, be aware that there is a waiting period once you place an order for a Swarm, so if you are sure you are going to go that way, the sooner you order the better.
To be clear, I would make sure I had the two RELs you have optimally placed before I do anything. If that gets you where you want to be, it costs you nothing. Problem solved at no cost.
Most of the time and in most rooms, it takes 4 subs to get the low frequencies where you would like. I suspect you will find that your room is not be an exception to that general rule. I think you would have a more cost effective and possibly better solution going with the Audiokinesis Swarm or the similar Debra than buying two more REL subs. Those things ain't cheap. If you get an AK Swarm, you won't need the existing RELs, so you might as well sell them and use the proceeds on offsetting the cost of the Swarm.
The Swarm system will go to a lower frequency than almost all of the REL subs will. I have nothing against REL subs, and use one in my living room system. But the distributed array approach is just so much better fundamentally that it is the way to go if you are looking to achieve the best results.
Again, before I spent any money, I'd send Duke my room dimensions and see what he thinks. He's seen the Swarm installed in a lot of different rooms, I'm sure. I have zero actual experience with a room like yours.