Calvert, Do ‘normal’ traffic conditions really exist? Why modelling variation & uncertainty is not a choice
- 1. Do ‘normal’ traffic conditions really exist? Why modelling variation & uncertainty is not a choice S.C. Calvert MSc TU Delft Masterclass May 2012 Challenge the future 1
- 2. Who am I? Simeon Calvert, MSc, 28 yrs • Traffic researcher at TNO & PhD-candidate at TU Delft • Graduated at TU Delft, Transport & Planning, 2010. • Specialisation in Traffic flow theory and traffic modelling • PhD-subject is on probabilistic traffic flow modelling. Challenge the future 2
- 3. Contents for today • Is considering ‘normal’ traffic conditions good enough? • Demonstration of variation in traffic flow • Focus of my research • Modelling variation using probability • Supervision Challenge the future 3
- 4. Is ‘normal’ traffic flow good enough? • Traffic is affected by day-to-day variations in traffic demand, weather conditions, road works, etc. • Normal practice: take an average/representative situation • Is it sufficient to consider normal traffic conditions when modelling traffic? And why? Challenge the future 4
- 5. Is ‘normal’ traffic flow good enough? • Demonstration of variations present on roads 8000 7000 7000 6000 Traffic demand: 6000 5000 5000 4000 4000 Congestion: 3000 3000 2000 2000 1000 1000 0 0 0 1 2 3 4 5 6 5 x 10 1.46 1.47 1.48 1.49 1.5 1.51 1.52 1.53 1.54 1.55 5 x 10 Challenge the future 5
- 6. Is ‘normal’ traffic flow good enough? • Capacity varied due to different weather conditions • Two scenarios with same input, but one varied and one the ‘average’ situation 1 0.9 0.8 0.7 Capacity factor 0.6 0.5 0.4 0.3 0.2 0.1 0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Cumulative probability source: Calvert, Taale, Snelder & Hoogendoorn (2012) Challenge the future 6
- 7. Is ‘normal’ traffic flow good enough? • Results: Scenario Median Travel times Average Travel times (minutes) (minutes) Variation in input 20.23 23.98 No variation in input 18.16 18.16 source: Calvert, Taale, Snelder & Hoogendoorn (2012) • Varied capacity leads to a (much) higher travel time! • Non-varied capacity does not sufficiently consider delays • NB: This is often the case, but not always! -Calvert & Taale (2012) Challenge the future 7
- 8. Focus of my research • Effects of (external) events on traffic flow • Such as weather, daily demand variation, incidents, … • Data analysis • Modelling variation in traffic • Models used for planning/forecasting & evaluation • Presentation of probabilistic model results • Especially for policy-makers ? Challenge the future 8
- 9. Modelling variation using probability • Traffic modelling -> macroscopic / microscopic Challenge the future 9
- 10. Modelling variation using probability • Advanced Monte Carlo • Many simulations, with a different combination of input values for each simulation • ‘Advanced’ refers to clever ways of selecting the input values Histogram of network delay (Systemtic sampling) 1 12 • INPUT: 0.9 OUTPUT: 10 0.8 0.7 8 Capacity factor 0.6 Frequency 0.5 6 0.4 4 0.3 0.2 2 0.1 0 0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 1 2 3 4 5 6 Cumulative probability Network delay (vehicle hours) 4 x 10 Challenge the future 10
- 11. Modelling variation using probability • Core probability • Variation is calculated in the core of the model as sets of probability distributions • Faster, completer, ☺ … but harder to implement Congestion if: K > Kcritical = 25 veh/km Challenge the future 11
- 12. Supervision • MSc: • 1 or 2 daily supervisors & professor • PhD: • 1 or 2 daily supervisors & promoter • High degree of independence expected Challenge the future 12
- 13. Supervision • Important for dealing with supervisors: • Make clear arrangements (about meetings, input, reporting, …) • Be independent, but keep regular contact with supervisor(s) • Remember: supervisors have been there before! • …but you do have a say how you progress. • Serving two masters (conflicting interests)* • Sort out problems quickly, don’t ignore them! Challenge the future 13
- 14. Opportunities / Interested? • MSc-thesis project internship • @TNO (or @ITS Edulab) • Email: simeon.calvert@tno.nl Challenge the future 14
- 15. Challenge the future 15