Rideshare systems can increase the efficiency of the transportation system of large metropolitan areas such as Los Angeles County. They do so by providing flexible and convenient commutes to passengers, thereby reducing the number of solo drivers. A real-time and practically applicable rideshare system needs to be designed with several conditions in mind. Drivers and passengers need to be matched quickly with minimal waiting time for passengers and drivers. In addition, drivers need to be provided with routes that will minimize their detours and the total traveling cost of the system. The rideshare matching (of drivers and passengers) and routing system also needs to be scalable to implement the system on a city-wide scale.
This project is informed by several key observations. First, dynamic ridesharing system performance can be improved by forecasting future demands from historical data. Also, the rideshare routes provided to drivers may need to be updated in real-time as traffic conditions change. Therefore, this project will develop a framework that will, in real-time, match drivers to passengers and route rideshare drivers that incorporates traffic data to provide improved solutions. The resulting solution framework will use commercial traffic simulation software to achieve this goal, in contrast with much of the existing literature which assumes static assignments that do not include traffic conditions. Utilizing common pickup and drop-off points for multiple passengers in this framework can potentially reduce travel costs and time. In addition, insights will be created into the adoption of rideshare among commuters through the analysis of individual rationality incentives. The framework will be scalable such that it can be applied to demand in large metropolitan areas.