It would only take 3,000 Uber cars to usurp regular cab services as the main form of transformation in New York City.

That's suggested by a new study conducted by the Massachusetts Institute of Technology, or MIT. New York City's 14,000 licensed fleet of cabs could be beaten out by a significantly smaller number, that is, if users are willing to share rides through carpooling services Uber and Lyft, or other companies offering similar services.

Carpooling Could Phase Out NYC's Taxi Fleet

In a study amounting to six months, a research team managed to create an algorithm which showed that 3,000 cars with four riders in tow could serve 98 percent of New York City's demand for a primary transportation network. What's more, average wait times for the study, at 2.3 minutes, stands far more amiable for the comparatively abject wait times of independently hailing a traditional cab in the city during peak hours.

The algorithm pulls data from 3 million taxi rides in the city, and works in real-time to "reroute cars based on incoming requests," said the researchers. Inactive cars are also sent off to high-demand areas, speeding up the service by 20 percent.

In comparison, ride-hailing services also have their proprietary algorithms for determining the dispatch of rides via their ride-sharing system, but these come with limitations, according to the researchers. One of the limitations is that "rider A" should be located within the same route as "rider B," and sometimes, all incoming request must have already been issued before the route is eventually created.

MIT's Algorithm

The algorithm created by MIT offers a defter complexity to implement ride-sharing, and it also gets better at learning the more it's used. The findings, simply put, predict that carpooling services could singlehandedly eliminate New York's iconic yellow cab, but despite this, it's not meant to imply that this is the trajectory New York City is in for.

Speaking to the Washington Post via phone, Daniela Rus, Director of MIT's Computer Science and Artificial Intelligence Laboratory, said that the study simply shows that a city's transportation network could still be functional even when reduced to 3,000 cars hovering about, provided that users are willing to share rides, of course.

In fact, the study can also be another way of improving the lives of taxi drivers, often exhausting themselves by working double shifts just to get by. Instead of half-a-day's shift, drivers "could work six- or eight-hour shifts," while cashing the same amount of profits, since "it's the same transportation need, it's the same level of payment that flows through the system."

Other Experiments

The study also found that the fleet could be lowered still. 2,000 vans with loaded with 10 riders could similarly serve 98 percent of the city's transportation network, with a slightly longer wait time of 2.8 minutes. MIT's algorithm dispatches the right type of vehicle in relation to the incoming request.

Furthermore, the algorithm best serves autonomous or driverless cars, since it automatically reroutes cars on real-time, as dictated by requests that filter in the system. If it could replicate its findings wholesale, MIT's algorithm could meld perfectly with autonomous ride-hailing services, which, by the looks of it, could soon be the future of commute.