Calibration and Validation of VISSIM Models_190909

CALIBRATION AND VALIDATION OF VISSIMMODELSVISSIM MODELS 

Peter Richardson  B.Sc., C.Eng., MICE, MIHTHollywood Consultancy, 

Woodford Green,Essex, 

England,  IG8  9LG

St Petersburg, September 2009Санкт–Петербург, сентября 2009

Email: [email protected]: www.hollywoodconsultancy.co.uk

CALIBRATION & VALIDATION: BASICSCALIBRATION  & VALIDATION: BASICS• Calibration is a process where the modeller selectsthe correct model parameters that cause the modelto reproduce the observed traffic conditions asaccurately as possibleaccurately as possible.

• Validation is a process where the modeller checksValidation is a process where the modeller checksthe overall model outputs for observed values oftraffic performance, e.g. traffic flows, travel times,speeds, queues and delays. This uses data not usedin the calibration process and it can be described asan independent check of the calibrated modelan independent check of the calibrated model.

CALIBRATIONCALIBRATIONModel calibration consists of:

• Network calibration

• Trip matrix development

• Trip assignmentTrip assignment

NETWORK CALIBRATIONNETWORK CALIBRATION• Network calibration ensures that vehicles travel 

l th h th t kproperly through the network. 

• Routes and junctions operate correctly i.e. as observed. 

• Queues need to be taken into account‐ long queues indicate unsatisfied demand i.e. overloading. In congested networks the real i.e. demand flows should be input, not the flows actually passed by the network.

NETWORK CALIBRATION• Nodes should be modelled correctly e.g. stoplines and give way points as existing.give way points as existing.

• Links should operate as observed e.g. lane changes.• Vehicles should have correct free flow headways• Vehicles should have correct free flow headways.• At signals lanes should have correct discharge rates/saturation flows and degrees of saturationrates/saturation flows and degrees of saturation.

• Typical discharge rate = 1800 vehs/hr.DOS t l fl *100/ i fl It h ld t• DOS = actual flow*100/maximum flow. It should not exceed 100%! V hi l h ld t l t t d i t• Vehicles should travel at correct speeds in correct lanes.

QUEUE FOR SATURATION FLOW SURVEYSQUEUE FOR SATURATION FLOW SURVEYS

CYCLES AT SIGNALSCYCLES AT SIGNALS

MOTOR CYCLES AT SIGNALSMOTOR CYCLES AT SIGNALS

LANE USAGELANE USAGE

TRIP MATRIX DEVELOPMENT• Trip matrix development ensures that the O‐D matrix (Origins and Destinations of traffic flows matrix) is as(Origins and Destinations of traffic flows matrix) is as accurate as possible. This is particularly important for Dynamic Assignment as opposed to Fixed RoutesDynamic Assignment, as opposed to Fixed Routes. 

• With Dynamic Assignment the model must iterate to find the best routesfind the best routes.

• Convergence criteria are used to ensure that the h d h bDynamic Assignment has converged to the best O‐D.

• With Fixed Routes the model is given sufficient information on a junction by junction basis to apportion all traffic flows.

TRIP ASSIGNMENTTRIP ASSIGNMENT 

• Trip assignment is how vehicles travel through• Trip assignment is how vehicles travel through networks. This involves route choices which depend on travel costs through the networkon travel costs through the network.

VALIDATIONVALIDATION • Typical validation checks are carried out on link flows 

d li fland screenline flows.

• Journey times should correlate. 

• There is no precise way of validating queue lengths because of the volatility of the observed data, i.e. it y ,is difficult to measure queues accurately with sufficient detail.

• Delays at junctions.

• Routes used• Routes used.

GEH VALUE

VALIDATIONVALIDATION • Can be difficult‐ random seeds are also used to simulate traffic variabilitysimulate traffic variability.

• A report should be prepared to describe how the d l h b lib d d lid dmodel has been calibrated and validated.

USEFUL REFERENCES

• Fellendorf M and P Vortisch “Validation of the MicroscopicFellendorf, M., and P. Vortisch,  Validation of the Microscopic Traffic Flow Model VISSIM in Different Real‐World Situations,” Paper presented at the Annual Meeting, TRB, Washington, DC, 2001.

• Volume III‐ FHWA Guidelines for applying traffic microsimulation modeling software, US Department of Transportation.

l 2 S i 2 i l f d & id• Volume 12, Section 2, Design Manual for Roads & Bridges, Highways Agency, London, UK.

• DTO Modelling G idelines Transport for London UK• DTO Modelling Guidelines, Transport for London, UK.