![8
Model Development
• Maps from street directories
• Link speeds from onboard
bus equipment
• Lines information from
public transport guides
• Demand matrices from
ticketing data
• Development data from
various agencies
[ez-link reader]
[IDFC console]](https://image.slidesharecdn.com/d4ba4669-a931-4022-903e-184357df244f-161101143012/85/PT-Modeling-in-S-pore-Terence-Ng-8-320.jpg)
![24
Select Link
• Ability to select line, nodes, links or a
combination of criteria
• For example: -
MW[1] =
SELECTLINK((L=12809-40025* + LINE=2400)
& (L=12200-40026* + LINE=2400))
for a new bus service
connecting 2 different MRT
stations](https://image.slidesharecdn.com/d4ba4669-a931-4022-903e-184357df244f-161101143012/85/PT-Modeling-in-S-pore-Terence-Ng-24-320.jpg)
PT Modeling in S'pore [Terence Ng]
- 1. Citilabs 2009 Asia-Pacific User Conference Public Transport Modeling in Singapore using TRIPS and CUBE Terence Ng Manager, Service Development SBS Transit Ltd Singapore 11 Mar 2009
- 2. 2 Presentation Outline • Some facts on Singapore • Model Design and Applications • Limitations and Challenges • Migration to CUBE and its benefits • Further development work
- 3. 3 Some facts on Singapore • Land area: 700 km2 • Population: 5 million • Over 50% use public transport • Daily Rides: – Bus (3.0 million) – Rail (1.6 million)
- 4. Public Transport in Singapore • Bus/MRT/LRT are main modes • 2 major multi-modal operators – SBS Transit – SMRT SBS Transit Bus Area Rail Lines SMRT Bus Area Rail Lines Central Area
- 5. 5 SBS Transit Bus • 2,800 buses • 254 routes • 2.3 million rides daily • 16 bus interchanges, 16 terminals • Over 3,000 bus stops Rail • 360,000 rides daily • 15 MRT stations, 19 LRT stations in operation
- 7. 7 Building up Modelling Expertise • Acquired TRIPS in 2001 • Evaluate impact of route changes • Assess viability of new route proposals • Test many options before determining the best proposal • Being self-sufficient in transport modelling.
- 8. 8 Model Development • Maps from street directories • Link speeds from onboard bus equipment • Lines information from public transport guides • Demand matrices from ticketing data • Development data from various agencies [ez-link reader] [IDFC console]
- 9. 9 Model Structure Demand Matrices • Origins and destinations from smart card data Service Information • Routes, headway, run times, capacity, • Mode Type, fares, operator Assignment • Time Period - AM Peak 0630–0830 hrs • User Classes - Adult Fares - Child/Student Fares - Student Pass • Mode characteristics - Boarding Penalties - Transfer Penalties - Walk Factors - Wait Factors - Value of Time Total Boarding and Passenger KM • By route, direction, mode and operator Load Profile • By route and direction at nodes/stops Stop-to-stop movements • By route and direction INPUT OUTPUTPROCESS Base Map • Nodes (Bus stops, MRT stations, interchanges, terminals) • Links between nodes
- 10. 10 Calibration • Total Boarding and Passenger KM by route and direction • Total passenger volume leaving towns • Station-to-Station movements for MRT • Heaviest load points by route
- 11. 11 Major Applications • Implementation of over 20 new routes and more than 50 route changes (2001–2008) • Commencement of North-East MRT Line and Sengkang LRT East Loop (2003) • Opening of Punggol LRT East Loop and Sengkang LRT West Loop (2005)
- 13. 13 Strictly Public Transport • Demand matrices built directly from smart card data • Does not account for effects on private transport modes
- 14. 14 Demand Matrix • Full details within SBS Transit areas • Aggregated data between SBS Transit and SMRT areas • No data within SMRT areas • SMRT Rail Captive zones need treatment • Difficult to manipulate in *.mat format Central Area
- 15. 15 Total Generalised Costs • Usually satisfied at global level and easy to skim • Does not sufficiently address the affected minority • More details needed at precinct and route level
- 16. 16 Limited Output from Reports • Planners need detailed breakdown of passenger impact in terms of fares, journey times and number of transfers for any service proposal • Too many skimming process slow down model run times • Detailed computations still done manually outside of model
- 17. Migration to CUBE and its benefits
- 18. Migration to CUBE • Upgraded to CUBE since Jul 2008 • Ease updating of network and matrices • Enhance evaluation of proposals • Automate generation of useful planning data • Better interface with own systems
- 19. 19 Program Control • Script vs Menu based
- 20. 20 Network Map • Use of layers • Easier to navigate and update interactively • Wider choice of colour sets
- 21. 21 Inputs in Database Formats • Nodes, links, matrix records can be maintained in DBF formats easily editable in Excel
- 22. 22 Assignment • Program boxes reduced significantly • Can put more functions in each program group
- 23. 23 Fare Matrices • Use of station-to-station fares • Better represent through-fares between rail lines
- 24. 24 Select Link • Ability to select line, nodes, links or a combination of criteria • For example: - MW[1] = SELECTLINK((L=12809-40025* + LINE=2400) & (L=12200-40026* + LINE=2400)) for a new bus service connecting 2 different MRT stations
- 25. 25 Reports • Transfers between operators, modes • Stop-to-stop movements • Path analysis
- 27. 27 Further Development Work • Sensitivity tests of assignment parameters and fare models • Improve quality of reports • Matrix estimation using screen line flows • Path analysis with through fares
- 28. QUESTIONS?