3 ms·
I don't think this post accurately represents traffic engineering or the nature of congestion. As far as I know, traffic engineers have been doing sophisticated
by 0xcde4c3db 11y ago
I don't think this post accurately represents traffic engineering or the nature of congestion. As far as I know, traffic engineers have been doing sophisticated dynamic simulations for decades. Likewise congestion is naturally more likely when a road is heavily utilized, but it's primarily a coordination problem rather than a capacity problem (cf. ALOHA vs. slotted ALOHA).
- simval 11y agoI'm the author of that piece. Just pointing out that I am a traffic engineer, I have seen, analyzed and criticized traffic studies. Believe me, your faith in traffic engineering is quite misplaced. Most cities do not have the resources to do complex dynamic simulations. In Québec, the Transport Ministry has a model (EMME) supposed to take into account trip choices and modal choices, but I haven't heard any confirmation of its prediction capabilities. The big issue with such simulations is the amount of input data you would need to adequately model reality. A metropolitan area has hundreds of thousands of workers, or even millions, to have a reliable model, you would need to know where everyone lives, where they work, what is their work schedule, their income level, etc... Obviously, collecting that data is essentially impossible, after you add up the uncertainty of the inputs, you often realize that the uncertainty makes the results completely unreliable, almost no better than an educated guess.
- 0xcde4c3db 11y agoFair enough. My thinking was that you'd use probabilistic methods and run many trials to see where the failures are, but I guess that's likely to fail if the real-world patterns are too cross-correlated, which they undoubtedly are (rush hour, etc.).