IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS, VOL. 16, NO. 3, JUNE 2015

Scanning the Issue and Beyond: Transportation

Games for Social Transportation

This issue brings us a few new ways and directions to evaluate the past ITS practices and look into the future applications of intelligent technology for intelligent transportation. Especially, our three review articles address related issues in urban ITS, smart rail systems, and mitigating adverse weather impacts to mobility. For my part, inspired by the current success of Uber-type taxi services, I will discuss the importance and emerging trend of transportation games for social transportation.

TRANSPORTATION GAMES FOR SOCIAL TRANSPORTATION

Recently, I read an article on Uber’s business model and its bidding mechanism for real-time matching of demands and supplies, which claims that this model has revolutionized customers’ riding experiences, released fragmentized values, reconstructed social networks for connection beyond mobility, and should win the Nobel Prize for Economics. Having never used the Uber service myself, I am unable to judge the article’s claim from my personal experience, but I believe Uber’s phenomenal growth from Dollar 200K to Dollar 40 B in value and from cabs to airplanes in vehicles, in just five years, must mean something significant. When discussing this with my students in game theory, they told me a few interesting stories of frauds for profits by Uber’s taxi drivers. However, I am still convinced that Uber’s practice, or in general, Transportation Games, will gain momentum shortly for fair and efficient mobility, and beyond, in this increasingly connected world. 

Through pervasive applications of smart phones, GPS, and other mobile devices, along with social media platforms such as WeChat, Facebook, and Twitter, individuals now can generate, obtain, and share real-time traffic information, and traffic operation centers can notify, direct, and interact with them and vehicles, together creating and altering traffic strategies via crowdsourcing, rather than the traditional fashion of command and control. To me, this offers us an exciting window of opportunity for various emerging live and grand Transportation Games in a new world, actually in three worlds: Physical World, Mental World, and Artificial World, in the sense of Karl Popper’s Worlds I, II, and III. Technically, I would like to call it the Era of Transportation in Cyber–Social–Physical (CSP) space, or the age of parallel transportation. Transportation games, or computational transportation games in CSP, to be more precise, will be the key to the success of ITS in this new age. This is an important issue that I should return to for a deep discussion later. 

Dynamic ridesharing service providers, such as Uber and Didi, provide primitive but powerful mechanisms and platforms for real transportation games in the real world. In just a few years, they have presented many interesting but complicated patterns of interactions and behaviors for mobility with various levels of social, economic, and engineering complexity. For example, The Washington Post reports that Uber’s dynamic price mechanism “Surge Pricing” actually “reduces demands for cars” since “less people want a car for a higher price” and “shift drivers to areas of high demand” instead of “create new supply,” as pointed out vividly by the Times, that Uber’s fee in Sydney “hikes upwards of four times their normal rate” after “an armed assailant burst into a city cafe and took hostages.” In China, the leading ridesharing company, Didi, has used a variant of Dutch auctions to match drivers and customers in the same area: a customer can set and modify her or his own price and a driver can grab his or her favored orders according to customers’ prices and destinations, creating a complicated game scenario among operators, drivers, and customers, a powerful display of smart phone supported social transportation games for business operations.

Transportation games have a long history in transportation research: Wardrop’s selfish routing games in 1952, Rosenthal’s congestion games in 1973, and Monderer and Shapley’s potential games for heavy-duty vehicles in 1996 are just a few examples. More recently, Stackelberg strategies have been widely employed and experimented for transportation and mobility, e.g., Krichene et al. in 2012 and Bayen in 2013 have studied Stackelberg routing strategies on parallel networks with horizontal queues for reducing traffic congestion, under the assumption that the central authority can incentivize the routes of a subset of players on a network and the remaining players choose their routes selfishly. Such “soft” traffic suggestions, rather than “hard” traffic controls, could be an ideal match with the total or complete traffic control approach I have proposed in the Editorial from Transactions on Intelligent Transportation Systems, Vol. 15, No. 2, 2014. 

Just before I am finishing this editorial, I sadly learn that John F. Nash and his wife were killed in a taxi crash in New Jersey. Nash, born in the year of 1928 when game theory was invented by John von Neumann, was a pioneer in games. Nash equilibrium has played a central role game research, and Nash strategy has been used widely in earlier transportation games. Personally I have a deep and special interest in game theory. Thirty years ago when writing my PhD dissertation on coordination theory of intelligent machines, I tried both machine learning and game-theoretic approach, with Nash, Cooperative, and Stackelberg strategies. As social networks and connected societies are becoming normal of our life, we should speed up our work in transportation games for social transportation, and make our ITS real smart, in human’s terms.

FEI-YUE WANG, Editor-in-Chief

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Scanning the Issue and Beyond： Transportation Games for Social Transportation.pdf