Taxis By Algorithm: Streamlining City Transport With Graph Theory
New submitter Mark Buchanan (3595113) writes with a story about research from scientists at MIT, Cornell and elsewhere showing "that big city taxi systems could be made 40% more efficient with device-enabled taxi sharing. We could cut miles driven, costs, and pollution with the right application of just data and algorithms, and do it while introducing no more than a 5 minute delay to any person's trip. " Letting such algorithms compete seems an excellent reason to encourage, rather than reject by law, ride-coordination services like Uber and Lyft.
I don't want to share my cab.
Letting such agorithms compete seems an excellent reason to encourage, rather than reject by law, ride-coordination services like Uber and Lyft.
Taxi licensing laws aren't about giving the CUSTOMERS good service. They're about limiting competition so the licensed cab owners have a regulated oligopoly that limits competition and keeps the prices higher than market-clearing.
It's much like the laws limiting car sales to dealers that are giving Tesla such a problem.
This is crony capitalism at its most blatant.
Bantam Dominique roosters crow a four-note song. Once you've heard it as "Happy BIRTHday" you can't NOT hear it that way
And making a big city taxi system more efficient helps the existing taxi companies how?
[I was beaten to this by an AC, but since many filter AC comments I'll jump in...] The Taxi companies have no interest in something that will effectively reduce their revenue.
In Turkey, I saw even better thing. Idea is this: public transport bus is too slow and awkward: stops are either sparse - lots of walking, or dense - making traveling too slow, and taxi for single person is too expensive (fuel + driver). In Turkey these is this "Dolmush" thing, which is mini-bus, that stops anywhere (like Taxi), costs fix rate (like public transport) and is just practical. It kicks ass of all other forms of public transport *AND* computerized car/taxi sharing.
839*929
There is a case to be made for taxi regulation. It protects passengers, which is really the main reason taxi regulation exists. In order to fund that regulation, they allow companies artificial monopolies.
The last thing you want is a totally unregulated taxi industry. There is a reason these kinds of things became regulated in the first place.
There was a pilot for this program 4 years ago in NYC:
http://www.nytimes.com/2010/02...
Also there was strike that mandated it 7 years ago for a few days.
http://www.nytimes.com/2007/09...
In short, no one liked it. If people wanted to have a delayed trip and people with them, they'd just take the Subway.
We have a dominant Taxi company in the city where I live, one so influential and powerful that it has landed a former city councilman in Federal court for accepting bribe money to stifle competition from other companies. They are, of course, bringing some serious legal pressure to bear on our politicians for allowing Uber to operate independent of codes regulating the taxi business. Seems that the biggest barriers to improving cab transportation are existing regulations and conflict with existing companies, not technical. In Dallas, the attempt to stifle Uber went so far as to include police stings of Uber drivers: http://www.dallasobserver.com/... Either existing taxi companies need to adopt new technologies like this on their own, or citizens have to demand restructuring of the rules governing them. It seems in many places that taxi companies view these more as a threat than a benefit.
The law is a cab is legally bound to take you anywhere in the five boroughs, whether they want to or not. Of course, if you tell them where you want to go before you get in, there's not much you can do to force the issue, except maybe getting their plate number and report it, which you probably won't do anyway. So savvy NYers don't give up the destination until they're in the cab.
Long story short, a system that requires you provide both the pickup AND arrival points will require some serious clampdowns to keep uptowners and outer-borough folk from being left out in the cold.
.
Prisencolinensinainciusol. Ol Rait!
40% more efficient for each individual getting the ride? Those are the people the companies exist to serve. Or 40% more efficient for the company? 40% more efficient for who?
As far as cabbies are concerned, the optimum algorithm will be whatever maximises their revenue. Any algorithm that doesn't will probably be vulnerable to cheating, i.e. a rogue cabbie that can make more money exploiting some aspect of the algorithm will do so.
There's even more to it especially in extreme downtowns. You mention that parking downtown is expensive - but picture if we managed to have a transportation system efficient enough that we didn't even need that much parking downtown. That parking garage could be another skyscraper. A few more of those and you might be able to justify some fancy people moving system like airport slideways, automated shuttles, etc...
You can even end up with a situation where an suburbanite like myself ends up walking more because of the trek from parking lot to building than the urbanites do, while never setting foot outside unless it's by choice.
I don't read AC A human right
My daughter and I do not want to share a cab with your drunk ass friend. No thanks.
On that note, I'd like to see an analysis this system on the economics of taxi operations. Shaving 40% off total miles driven would seem to have a significant impact on a cabby's income unless the sharers are charged the same as they would be otherwise. Additionally, if there's a delay, who pays for the time and mileage?
Sure there's a positive environmental impact, but it seems as if there's an additional cost born by someone, even though the miles driven goes down. Both fairness and fareness demand some explanation here.
I had a friend tell me about the time he filled in for a day as a dispatcher for a cab company. He said he tried to do it efficiently by looking at where the customer was and where the closest non-busy taxi was and sending that one. Turns our the taxi drivers didn't like it. They wanted him to send them out in order so they all got the same number of fares.
Coder's Stone: The programming language quick ref for iPad
I'd extend this idea to all cars and traffic lights. If cars could talk to traffic lights and each other, that'd save tons of time, miles driven, costs, and pollution. Such a network can optimize car routes to choose less congested ones, recommend car speeds, and change traffic lights accordingly to provide green light corridors.
Split all the fairs evenly to all drivers - since fairs are decided by millage anyway it shouldn't be a big deal... unless they figure it out and realize fewer miles are being driven with a efficient system. If they don't charge for the distance to the pick up, then that factor would be a lower overhead cost and save them money.
Tips. Well, that is not actually randomly distributed so I could see complaints about not getting more time around certain areas at certain times. They won't ever agree to pool tips.
How about you just save up as a company and replace all the humans with robots in a decade.
Democracy Now! - uncensored, anti-establishment news