Research and Application of Key Technologies for Request Prediction and Assignment on Ridesharing Platforms

Abstract

DiDi ridesharing platform provides users with accurate and high-precision travel services in real-time by using data-driven technology, machine learning method, large-scale distributed computing, operations optimization and other technique. The platform has made great breakthroughs in the key technologies of intelligent prediction and dispatch of travel platforms: Estimated Time of Arrival (ETA), Intelligent Dispatching and Supply and Demand Forecasting. We proposed a novel deep learning solution to predict the vehicle travel time based on floating-car data. We also present an order dispatch algorithm in large-scale on-demand ride-hailing platforms. While traditional order dispatch approaches usually focus on immediate customer satisfaction, the proposed algorithm is designed to provide a more efficient way to optimize resource utilization and user experience in a global and more farsighted view. We deploy the spatiotemporal multi-graph convolution network (ST-MGCN), a novel deep learning model for demand forecasting.