Optimizing ride-hailing systems
Sean Qian aims to leverage the platform of ride-hailing companies to benefit everyone in the transportation system: reduced travel times and costs for travelers, more profit for service providers, cost-effective methods for government agencies, and less emissions for our environment.
Imagine that you’re trying to get back home at around 5 p.m. on a Tuesday. Your car is moving like a snail, and you start to wonder where everyone else is going—heading home like you, meeting friends for dinner, or going to a movie? While these may be true, some drivers are simply wandering around to find customers or standing by forever at curbs.
Many challenges arise as more people choose to call Uber or Lyft instead of taking public transportation. These ride-hailing companies aim to make people’s lives more convenient and reduce traffic; however, the outcome is often actually more congestion. Sean Qian has identified two main problems with the growth of ride-hailing.
The first one is related to driver habits. When the vehicles are not taking passengers, their drivers tend to cruise around to find the next order, which substantially increases congestion. The other issue is about the usage of public spaces, namely curbs. Curbs are designed for the public to share. However, a lot of them have become popular pick-up and drop-off locations for ride-hailing companies, which causes inconvenience for other drivers and triggers safety concerns.
“And now the cities are asking, ‘how can we regulate that?’” said Qian, an associate professor of civil and environmental engineering.
To answer this question, Qian is developing a cyber-physical-social system that lets cities leverage the platform of ride-hailing companies to provide benefits for everyone in the transportation system. Qian will collaborate with the University of Pittsburgh and the Department of Mobility and Infrastructure (DOMI) of the City of Pittsburgh. Their research is funded by the National Science Foundation (NSF).
We’re going to use optimization techniques, simulation, and AI to make the best decision for the city.Sean Qian, Associate Professor, Civil and Environmental Engineering
Their first step is identifying the current needs of all parties involved. We travelers obviously wish to reduce our travel costs and time. The ride-hailing companies want to increase their revenue, and the government aims to reduce congestion, energy use, and emissions in the city.
To benefit all parties, Qian’s team aims to leverage the ride-hailing platforms and build a new business model through the Public-Private Partnership (PPP).
In this model, the government agency regulates the market and sets up specific rules for ride-hailing companies. Take crowded streets during peak hours as an example: the companies may have to pay a surcharge if they wish to ride on those streets. To address the public space problem, the government agency can charge these companies if they want to drop off or pick up passengers from the most popular roads or curb spaces.
People respond to incentives. Currently, ride-hailing drivers have total freedom to choose their routes. This is what researchers call “selfish routing,” meaning each driver will take the route most advantageous to themselves, regardless of other drivers. To encourage ride-hailing companies to route the system optimally, government agencies could give them subsidies in return for a guaranteed travel congestion reduction.
Meanwhile, the researchers plan to build incentives into the ride-hailing fare systems to encourage travelers to change their departure time and routes. For example, riders could receive credits or discounts if they agree to wait during peak hours and place the order slightly later.
This system is particularly appealing to cities because it overcomes the barriers to implementing such incentives in the past. Traditional incentives, such as congestion pricing, are expensive to operate and can lead to social equity issues. With this new incentive, however, public agencies can save more money, and travelers can use ride-hailing services voluntarily.
In this case, the key to success lies in understanding how people behave. Qian’s team aims to combine machine learning techniques and existing data about travel congestion during the day, the way people take buses, and ride-hailing drivers’ vehicle routes, to learn more about people’s current choices and predict their reactions to various types of incentive.
“We’re going to use optimization techniques, simulation, and AI to make the best decision for the city,” said Qian. “Basically, we want to tell the city and ride-hailing services ‘this is how much you should charge so that the whole system can be run most effectively for both of you.’”
Overall, Qian’s team envisions a win-win-win outcome for everyone involved in the ride-hailing business: reduced travel times and costs for travelers, more profit for service providers, cost-effective methods for government agencies, and last but not least, less energy use and emissions for our environment.