Perhaps you have never been to New York City and witnessed the 24 hour a day welter of taxis that crisscross the five boroughs endlessly. The vast majority of them are old Ford Crown Victorias with about a bazillion miles on them. These are not cars known for high fuel economy or low carbon emissions. A new study by MIT proves mathematically that carpooling is more efficient, causes less pollution, and costs less than traditional taxi fleets.
Taking data from over 3 million taxi rides in The Big Apple, MIT’s CSAIL computer science lab found that just 3,000 vehicles from services like UberPOOL and Lyft Line could replace New York Cities entire fleet of 14,000 taxis. Even better, carpooling would reduce congestion by 300%. Travel times would be hardly affected, with the average wait for a ride calculated at 2.7 minutes.
The MIT mathematicians created their own algorithm. It is more efficient than the ones used by Uber or Lyft, which often require a trip to be fully booked before a route is created. The MIT model works in real time to re-route cars based on incoming requests. It even sends idle vehicles to high demand areas, which can speed up service by 20%.
“A key challenge was to develop a real-time solution that considers the thousands of vehicles and requests at once,” says CSAIL Professor Daniela Rus. The system does that by creating a graph of all vehicles and requests, then calculating every possible trip combination to determine the best assignments. Any vehicles without trips are rebalanced and sent to high demand areas.
The system could cover 98% of trips in the Big Apple using 3,000 four-passenger vehicles and 95% of trips if 2,000 10 passenger vans are used. What’s more, the algorithms would get better the more they’re used, leading to even greater efficiency. “A system like this could allow drivers to work shorter shifts, while also creating less traffic, cleaner air and shorter, less stressful commutes,” Rus says.
Earlier this year, Elon Musk revealed that Tesla Motors is working on just such a ten passenger vehicle for urban ride sharing duty. It will be based on the Model X and Musk says a concept will be shown publicly in 2017.
As terrific as MIT’s new mathematical formulas are, New York and other cities are not looking for ways to do away with their taxi fleets. The point of the study, though, is to look toward a more efficient future. As much as Uber, Lyft and other services have changed urban transport in the last few years, autonomous and electric cars could shake things up even more. “It’s important that we as researchers do everything we can to explore ways to make these transportation systems as efficient and reliable as possible,” says Rus.