利用大数据对微观模拟ABM进行校准.pdf

1、1|WWW.BENTLEY.COM|2023 Bentley Systems,Incorporated 2023 Bentley Systems,IncorporatedUtilizing big data for calibration of a microsimulation ABMGaurav Vyas,Peter Vovsha,Joe Carr-Bentley SystemsPresented at the TRB Innovations in Travel Analysis and Planning Conference,June 4-6,20232|WWW.BENTLEY.COM|

2、2023 Bentley Systems,Incorporated Common Data Sources in Model Development Household travel survey(HHTS)data Main source of disaggregate estimation But large-scale HHTS is expensive and difficult to recruit Even 3,000 5,000 households are becoming increasingly problematic Other data sources include

3、Traffic counts Transit ridership(APC,e-ticketing/smartcard,on-board survey)Primarily only used for model validation and manual calibration3|WWW.BENTLEY.COM|2023 Bentley Systems,Incorporated ProsBecoming increasingly available from vendors“Big data”trip tables can be used to support aggregate 4-step

4、models in practiceConsNot behavioral(no details about trip purposes or individual attributes)No person ID to identify individual activity patternsGap between aggregate data structure and disaggregate ABM structureReality of transportation industry“Big data”as a replacement?Minimize the need of large

5、-scale HTS and take advantage of new types of dataCentral questionHow can big data or traffic counts be effectively used for the simultaneous calibration of travel demand model systems in addition to other data sources?4|WWW.BENTLEY.COM|2023 Bentley Systems,Incorporated Known Limitations of conventi

6、onal estimationSub-models are estimated one at a time based on the“perfect”observed prior choices However,they are applied conditional upon each otherLOS i.e.,accessibilities are assumed fixed and external However,in model application LOS is equilibrated and it deviates from the LOS used in model es