The FLIP-FLOP project (Flexible Line On-Demand Public Transport) is developing an AI-powered tool to simulate a more sustainable and efficient mobility service that combines the advantages of traditional bus routes with the flexibility of on-demand shared taxis. The service operates similarly to a regular bus line but can adapt flexibly to passenger stop requests. By dynamically adjusting vehicle usage and route planning in real time, the project aims to ensure efficient and demand-driven transportation – offering high capacity during peak hours and flexible service during off-peak times.
AI-supported optimization for seamless integration
The new combined mobility model utilizes various AI-powered methods:
- Generative neural networks enhance traffic demand forecasting by enriching synthetic populations and integrating additional data sources, such as OD matrices from existing traffic models or traffic information from mobile network data.
- Spatiotemporal machine learning models link historical and real-time data to accurately predict travel times.
- Heuristic and exact optimization techniques are combined with reinforcement learning for strategic and operational planning.
These algorithms determine vehicle types, station plans, and on-demand operations based on real-time demand forecasts.
Dynamic adaptation and alternative propulsion systems
The FLIP-FLOP mobility project simulates an innovative mobility service where passengers can book rides either in advance or spontaneously on-demand. The public transport system is designed to adapt in real time to each new demand to ensure optimal service quality.
Additionally, the project explores methods for integrating alternative propulsion systems, such as electric and hydrogen vehicles, as well as advanced technologies like automated vehicles. The efficiency and benefits of these innovations are systematically evaluated.
