- Real-time customer experience
- Social Event Analytics
- Bus Bunching
- Public Transport Awareness
- Dynamic Optimisation of City Intermodal Transportation
- Analytics and optimisation for more inclusive transport
- Denoising Infrastrcucture Data for Intelligent Transport
- Dynamic Intermodal journey Advisor
Smarter Urban Dynamics - Dynamic Optimisation of City Intermodal Transportation
Increasingly available sensor data allows us to view an urban transportation network as a "living" system, with sensing capabilities and a great potential to develop self-adaptation capabilities, as well as smarter communication with the users of a transport network.
In real life, transportation networks feature uncertainty about their current and predicted status, including uncertain bus arrival times, and unforeseen events such as accidents, traffic jams, delays, and bus or train breakdowns.
Dynamic Optimisation for City Intermodal Transportation, or DOCIT, is a first-of-a-kind project at IBM Research. The theme of the project involves optimising a multi-modal transportation network, and improving the way people make use of the network. A key aspect of our work is the ability to handle uncertainty and to build solutions that are robust to uncertainty.
In network optimisation, topics of interest include offline, strategic optimisation and online responses to disruptions, assisting an operator in the decision making process. Besides the operators, travelers are another key entity to consider. Unlike existing solutions on the market, we provide journey plans with risk hedging capabilities. Our planning engine uses a plan quality metric that quantifies the robustness of a plan to uncertainty, and provides plans optimised with respect to the robustness metric. Plans can include alternate, contingent options, providing a way to continue a trip even when an expected itinerary is invalidated due to uncertain conditions. This is complemented with tracking users and re-planning when needed, for a safer and better traveling experience.
Docit is a research intensive project. Expertise areas applied to Docit include AI planning under uncertainty, re-planning, optimisation, statistics and software architecture design.
Docit will be demonstrated with the city of Montpellier.
Contingent versus Deterministic Plans in Multi-Modal Journey Planning, Adi Botea, Stefano Braghin, Proceedings of ICAPS 2015, AAAI Press