About
Smart PT in a nutshell
Why do we need a Smart and Adaptive Public Transport System?
- Our transportation system is polarized: Either fully private (cars/taxis) or fully public and regulated (PT). Intermediate modes are hardly significant (mostly access/egress to other PT)
- PT services accommodate the demand of the past. Routes and timetables are mostly static, and services have difficulty to adapt to gradual changes in demand
- Demand- Supply mismatch brings PT to an ineffective equilibrium where PT users are mostly captives - the young, the elderly and low income households.
- Supplying more and more information on PT does not influence car users behaviour.
What is our approach to solve this problem?
- Systemically analyse the discrepancy between end-user travel demand and existing urban transit supply;
- Develop core algorithms of seamless adaptation of the inflexible transit supply to the current population activities and further PT evolution following up the gradually changing urban population patterns;
- Evaluate these core algorithms with a spatiotemporally -explicit high resolution agent-based simulation model using both synthetic and real historical data of urban dynamics
- Develop the pathways and policies for implementing SMART-PT approach in case studies of the project teams;
- investigate the potential impacts of this approach on end-users, service providers and regulators.
How will SMART-PT work?
The key idea is that we pull the public transport service "by the horns" to move it from its current state to an optimal state. The first stage is to understand in detail where and when people want to travel. This is done by mining the data from ICT sources – namely mobile phone call data records (CDRs). The next step involves creating and recognizing the flows of passengers and anticipating the possible changes in the spatiotemporal distribution of the demand in the future. Once demand projections are conducted we formulate the operational alternatives to accommodate the demand. This is done via genetic algorithms in agent-based simulations which try to find a solution that survives over time. Only good solutions will be adopted by the agent-passengers.
The model then evaluates and assesses the suggested alterations to the public transport network – are these small changes like switching between vehicles or timetable adaptations or are these big changes that require rerouting the lines of the service and changing mode of travel. The final result is a set of changes for implementation whereby decreases in demand will reduce the level of service e.g. change from regular bus line to paratransit or demand responsive taxi while increases in demand will bring about intensification of the services. The cycle then begins a new in the next time period (e.g. after 6 months or so).
What are the project outputs?
- SmartPT Interface: a system for collecting and analysing real or synthetic data on land-use/activity dynamics and users flows and providing spatiotemporal forecasts of demand
- SmartPT Kernel: a set of core algorithms that adapt PT operations to user demand by substituting modes and varying their routes and schedules
- SmartPT Sim: a spatially explicit high-resolution Agent-Based simulation environment that enables validation and sensitivity analyses of the SmartPT functioning in the city
- SmartPT Policy: a set of implementable policy measures and a participatory gaming platform for the seamless transition from inflexible to adaptive PT and further evolution of the transit system following evolving urban population pattern and end-user activities.
Where will the project take place?
The research framework will be employed for developing a proof-of-concept and investigate its effectiveness in case studies in Tel-Aviv, Stockholm and Leuven .
