US 12,260,358 B1
Using geocoded provider models to improve efficiency of a transportation matching system
Gaurav Gupta, Burlingame, CA (US); Ehud Milo, San Mateo, CA (US); Omar Khalid, Redmond, WA (US); Amy J. Kim, San Francisco, CA (US); Jacky Yi Han Lu, San Mateo, CA (US); Richard Zhao, San Francisco, CA (US); Robert A. Farmer, San Francisco, CA (US); and Akash Gaurav Shah, Millbrae, CA (US)
Assigned to Lyft, Inc., San Francisco, CA (US)
Filed by Lyft, Inc., San Francisco, CA (US)
Filed on Oct. 21, 2022, as Appl. No. 18/048,545.
Application 18/048,545 is a continuation of application No. 16/780,712, filed on Feb. 3, 2020, granted, now 11,494,714.
Application 16/780,712 is a continuation of application No. 16/125,563, filed on Sep. 7, 2018, granted, now 10,552,773, issued on Feb. 4, 2020.
Int. Cl. G06Q 10/0631 (2023.01); G06N 20/00 (2019.01)
CPC G06Q 10/06311 (2013.01) [G06N 20/00 (2019.01); G06Q 2240/00 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A computer-implemented method comprising:
identifying provider computing devices in a geocoded region comprising a plurality of geocoded areas;
determining estimated times of arrival for provider computing devices in the geocoded region based on GPS location information of the provider computing devices;
generating, utilizing a provider allocation model, a transportation flow matrix for the geocoded region based, at least in part, on the estimated times of arrival for the provider computing devices in the geocoded region;
determining, utilizing the transportation flow matrix, provider allocation shortages for at least one of the plurality of geocoded areas of the geocoded region;
generating, utilizing personalized provider behavioral models corresponding to the provider computing devices, and based on the provider allocation shortages determined utilizing the transportation flow matrix and on the estimated times of arrival, relocation prediction probabilities indicating individual probabilities that the provider computing devices will relocate to a target geocoded area of the plurality of geocoded areas;
selecting a provider computing device of the provider computing devices to relocate to the target geocoded area based on the relocation prediction probabilities generated utilizing the personalized provider behavioral models; and
providing, in real-time, to the provider computing device, a customized interface to guide the provider computing device to the target geocoded area with an updated estimated time of arrival for the provider computing device.