Batched Market Matching
In class we learned about matching markets and how they are typically split up into “buyers” and “sellers”. We learned about constricted sets and how sometimes not everyone can get what they want. This made me think about ridesharing apps such as Uber and Lyft. Millions of people all around the world use these apps to get around. Somehow both of these apps have a system that leads to perfect matchings. So, how does this work?
For ridesharing a bipartite graph would be set up such that people looking for rides are “buyers” and those giving rides are “sellers”. The first things to consider are the location and distance. Apps such as Uber use batched matching, which basically groups together all the “buyers” and “sellers” in the area. They want to optimize the average wait time for everyone in this area. This is the same concept as what we discussed in class relating to optimal assignments. Before this system, there was a first come first serve basis that would cater toward individual needs while increasing the overall travel time of the area. However, the batched matching system works to decrease the travel time creating a socially optimal outcome. In the case of Uber, time is basically equal to money so we can use the wait times to find market clearing prices. Well, this seems simple enough it is essentially matching markets but, what about Uber pool?
For Uber pool the price of a ride is split between the passengers. In order for this system to work the passengers would need to all be lined up such that the driver doesn’t have to change route to pick one of them up. The idea behind this is such that the driver will never have to stop driving and that the driver will continue to earn money. This is essentially the same thing a bus does except with optimized routes catered to the passengers. The bipartite graph of Uber pool would look basically the same as the previous graph except that the “sellers” would have multiple open spots that could be filled. Once again through batching the passengers would be matched optimally to a specific driver.
This system works to reduce the overall wait time of the passengers and increase the earning time of the drivers. It makes transportation more accessible and is more environmentally friendly. So overall using a batched matching system is good for the economy and for the environment.
Source: Uber Marketplace Matching | Uber