> I would be interested to see how the use of lots of small street routes around a main route affects the flow of the main route, and what the efficiency gain generally is.

I wouldn't assume there's an overall efficiency gain. That's part of the problem: in many cases, a shortcut that would work if a few cars used it causes problems (jammed intersections, etc.) if too many cars try to take advantage.

Someone who knows graph theory more than I do could comment, but more generally I don't think there's any guarantee that a situation where every vehicle uses a greedy algorithm will result in overall greatest efficiency. I could swear there was an article on this within the last year but I cannot find it.