—– excerpt below —–
New York study shows 13,000 vehicles could be slashed to just 3,000
by: Clive Cookson, Science Editor
New York City has more than 13,000 licensed cabs. Yet just 3,000 ride-sharing taxis controlled by an efficient citywide computer system could serve New York equally well, according to a study by Massachusetts Institute of Technology.
The analysis showed that 3,000 four-passenger cars could satisfy 98 per cent of the city’s demand, with an average waiting time of 2.7 minutes. While those trips would be delayed by an average 2.3 minutes compared with today because cars may take a less direct route to pick people up, most journeys would be faster than today because of the cut in congestion from fewer taxis on the road.
“Ride-sharing services have enormous potential for positive societal impact with respect to congestion, pollution and energy consumption,” said Daniela Rus, director of the computer science and artificial intelligence laboratory at MIT. “It’s important that we researchers do everything we can to explore ways to make these transportation systems as efficient and reliable as possible.”
Although carpooling and ride-sharing are not new, a combination of smartphones and computer routing has only recently enabled companies such as Uber and Lyft to offer these services efficiently on a large scale. While their algorithms are trade secrets, Professor Rus said the MIT system would be openly available.
The MIT team chose New York because it is a large city whose taxi movements are stored in a publicly available database. But the idea that a computerised ride-sharing system could produce huge savings would also apply to other urban areas.
The study, published in Proceedings of the National Academy of Sciences, was based on 3m taxi rides in New York during the course of a typical week.The data were used to construct an algorithm to route ride-sharing cabs as efficiently as possible through the city. It works in real time to redirect cars as requests come in, and proactively sends idle cars to areas where demand was expected to be high.
The study looked at ride-sharing based on different vehicle sizes, from small two-passenger cars to 10-person vehicles, but found a fleet composed predominantly of four-passenger cars worked best in New York.
“To our knowledge, this is the first time that scientists have been able to experimentally quantify the trade-off between fleet size, capacity, waiting time, travel delay and operational costs for a range of vehicles, from taxis to vans and shuttles,” said Prof Rus. “What is more, the system is particularly suited to autonomous [self-driving] cars, since it can continuously reroute vehicles based on real-time requests.”
The research algorithm could be applied in practice in various ways. Individual taxi companies or groups of companies could adopt it, or local government could compel taxi operators to participate in a shared citywide system.
Copyright The Financial Times Limited 2017. All rights reserved. You may share using our article tools. Please don’t cut articles from FT.com and redistribute by email or post to the web.