well i was thinking the same thing.
Since in nature they all grow together in soil, why not throw the soil bacteria onto a petri plate and see which form colonies (not all 5000 bacteria will grow under laboratory conditions. only about .5-1% of all known bacteria are culturable under laboratory conditions). Then take those colonies and put them in some type of defined liquid culture (separately, one species/strain per culture) and let them grow there. Then spin those cultures down and take the supernatant from each culture and set it aside for use later.
Then take P. aeruginosa (assuming you know which genes cause virulence) and make a mutant strain of it by latching a reporter such as GFP to one or more of those virulence genes that are controlled by quorum sensing. Then let the bacteria grow until in reaches the correct population density for quorum sensing to take effect and at that point, you should see some Green under UV light if you were to look at the colonies/biofilm/whatever (because at this point, the bacteria would be turning on those virulence genes you attached GFP to).
At this point, you can introduce the supernatants collected from step one (one by one into separate mutant P. aeruginosa cultures), and if they contain a quorum sensing inhibitor, you should see a gradual reduction in GFP (green light under UV) until it's all gone. This will tell you whether or not the supernatant contained a inhibitor and you can trace it back to the culture you got it from.