Thursday, November 30, 2006

Road Skill

This is a trivial problem, but its one that I like to think about. For some reason, it soothes me.

Here it is.

Take your average two or three lane highway. Think about it as traffic is streaming down the lanes. At some point, the average speed begins to drop. Cars become more tightly packed; the spaces between cars lessens. Every so often, traffic comes to a halt, or as close as makes no difference -- and then it starts up again. Sometimes it slows back down, other times it opens right up again.

How would you model that kind of changing activity? How detailed would the model have to be? How would you cause the model to accurately reflect how people actually drive, and how would you manipulate the model to reflect how they would drive in circumstances that aren't easy to reproduce -- rain, say, or a blocked lane?

I like thinking about stuff like that. I'm not knowledgeable about it (though I'd like to know more, I don't know of any traffic engineering journals aimed at the public), but I like thinking about it. I think it's fun.

No comments: