TU Delft and CTU win Taxify Self-Driving Fleet Optimisation Challenge

An international team from Delft University of Technology and the Czech Technical University in Prague have won the Taxify Self-Driving Fleet Optimisation Challenge. The challenge was to optimise the operation of the fleet of passenger service vehicles. The competition teams had to design a dispatching algorithm for a simulated fleet of autonomous taxi vehicles in the Estonian capital of Tallinn. A successful algorithm had to maximise reward for transported passengers, while respecting various constraints that simulate actual operating conditions. The winning solution was awarded a prize of € 4,000.

Taxify provides transportation across the city and ensures that a car is delivered to any place where Taxify operates within a few minutes. In the future, it plans to use autonomous cars. In solving this year’s challenge, the contestants had to assume that in 2022 this model of transportation would work in Tallinn, for which they had to program algorithms for the efficient management of such a fleet.

One of the challenges in the competition was to optimise the time spent between driving passengers and waiting for the next ride. It required the participants to develop various dispatching, routing, rebalancing, depot planning, and pooling algorithms. The future will reveal if the algorithms and ideas of the winners will once control the real self-driving fleets in our cities.

The winning team

The Autonomous Multi-Robots Lab (AMR) of Delft University of Technology and the Artificial Intelligence Center (AIC) from the Czech Technical University in Prague have long been working on the analysis and optimisation of transport systems. “The participation in the competition gave us a chance to apply our expertise in artificial intelligence to a challenging real-world problem” says Martin Schaefer, a researcher from AIC. "This competition was a great opportunity to combine the expertise of CTU in Prague and TU Delft in fleet management algorithms to solve a real world mobility problem and demonstrate the potential of Mobility-as-a-Service in Tallin" say Michal Čáp and Javier Alonso-Mora, researchers at TU Delft.

More information on research on ride-sharing on the Autonomous Multi-Robots Lab (AMR TU Delft) website: http://www.alonsomora.com/res/res_ridesharing.html