In a busy city, where empty taxis effortlessly find passengers without any hassle, a new system is setting the pace. The Sequential Route Recommender system aids taxis in selecting optimal routes to pick up riders. Professor Junming Liu, from the Department of Information Systems, along with co-authors, has introduced a system using a smart framework that assists taxis in determining the best roads to maximising earnings from their next passenger. By analysing the demand for rides and the availability of taxis, drivers can position themselves effectively at the right place and right time.
What makes this system special? It utilises technology to forecast potential passenger locations. By combining two models – a "graph convolution network" focusing on passenger locations and a "long short-term memory model" examining temporal trends – the system provides accurate predictions. This empowers taxis to assess the likelihood of finding passengers on various streets.
Armed with these real-time updates, drivers can make smart decisions about their routes. The system quickly calculates the best options, allowing drivers to reach passengers more efficiently. This approach enhances drivers' operational efficiently and reduces waiting times for riders.
Tests conducted using GPS data from taxis in Beijing demonstrate the effectiveness and efficiency of this new system. With this innovative technology, the future of taxi services appears promising, simplifying the process for individuals to secure a ride precisely when needed.
Liu, Junming; Teng, Mingfei; Chen, Weiwei; Xiong, Hui. "A Cost-Effective Sequential Route Recommender System for Taxi Drivers." September 2023; In: INFORMS Journal on Computing, Vol 35, Issue 5, pp. 1098-1119
Reference:
https://www.scopus.com/record/display.uri?eid=2-s2.0-85175475387&origin=recordpage