A Better Match for Drivers and Riders: Reinforcement Learning at Lyft
Xabi Azagirre, Akshay Balwally, Guillaume Candeli, Nicholas Chamandy, Benjamin Han, Alona King, Hyungjun Lee, Martin Lončarić, Sébastien Martin, Vijay Narasiman, Zhiwei Qin, Baptiste Richard, Sara Smoot, Sean V. Taylor, Garrett van Ryzin, Di Wu, F. Richard Yu, Alex Zamoshchin
Abstract
We used reinforcement learning to improve how Lyft matches drivers and riders. The change was implemented globally and led to more than $30 million per year in incremental driver revenue.
Topics & Concepts
Reinforcement learningReinforcementAeronauticsComputer scienceTransport engineeringArtificial intelligenceEngineeringPsychologySocial psychologyTransportation and Mobility InnovationsAutonomous Vehicle Technology and SafetyTraffic control and management