Agent-based model of greater Jakarta; current, and future work - - PowerPoint PPT Presentation

Agent-based model of greater Jakarta; current, and future work

Anugrah Ilahi Supervisor : Prof. K.W. Axhausen IVT ETH Zürich R4D colloquium , Zurich

Introduction

• Develop agent-based model in Greater Jakarta using MATSim
• Integrated choice model and agent-based model
• Explore the accessibility by public transport

2

Jakarta Now

3

Jakarta Now

4

MATSim loop

5

MATSim loop integrated with mode-choice model

6

x

Mode-Choice Model More info at: www.eqasim.org

Greater Jakarta Scenario Synthesis

7

Two parts:

• Supply (road network, public transport services)
• Demand (synthetic population)

Creating Public Transport Network; OSM & GTFS

8

Open Street Map

• 472 thousand nodes
• 1.2 million links

Multimodal Transport

• BRT
• Buses
• Rail
• Microbus (Angkot)
• Car
• Motorcycle

Synthetic Population (1)

• Population is synthesized using the data gathered by the Japan

International Cooperation Agency (JICA) in 2009 (JICA, 2009), which includes 3% of Households

• Synthetic agents are synthesized using Bayesian Network (BN)

and Generalized Raking (GR)

9

Synthetic Population (2)

10

Household data Individual data Geocoding location Data Joint/Fusion Population Synthesis using BN approach Fitting to marginal target using GR Marginal data Region by Age Marginal data Region by Gender Output Dataset of all variables needed

Choice Modelling (Parameter estimates)

11

Mode Parameters Estimate Public Transport

β numberOfTransfers, PT.

• 0.170

β inVehicleTime, PT [min-1]

• 0.120

β transferTime, PT [min-1]

• 0.048

β accessEgressTime, PT [min-1]

• 0.080

Car

α Car

1.227

β travelTime, Car [min-1]

• 0.066

Motorcycle

α MC

1.227

β travelTime, MC [min-1]

• 0.100

Walk

α walk

1.430

β travelTime,walk [min-1]

• 0.141

Other

β Cost

• 0.030

Calibration

θ parkingSearchTime, Car [min]

4.000

θ accessEgressTime, Car [min]

4.000

θ accessEgressTime,MC [min]

2.000

Activities simulated in MATSim (1)

12

Education Working Home

Activities simulated in MATSim (2)

13

Mode shares

14

30 26.8 41.6 8.6 31 24.1 40.5 4.4

5 10 15 20 25 30 35 40 45

Non-Motorized Transport Public Tranport Motorcycle Car MATSim JICA

Traffic Count comparison: MC on Thamrin St.

15

500 1000 1500 2000 2500 3000 3500 4000 4500 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 MATSim Counting

Traffic Count comparison: Cars on Thamrin St.

16

1000 2000 3000 4000 5000 6000 7000 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Trip Distance Band

17

Distance Trips Trips [%] Less than 1 km 105,372 7.72 1 - 5 km 514,413 37.67 5 - 10 km 300,938 22.04 10 - 20 km 248,785 18.22 20 - 30 km 95,119 6.97 More than 30 km 100,813 7.38

Accessibility

Conclusions

19

• We use a novel approach that integrates choice modelling in MATSim
• There is small difference of mode share between JICA and MATSim

model

• Number of motorcycles and cars in the MATSim model is higher than
• bserved.
• The accessibility is higher in the Greater Jakarta, and less so in

agglomeration cities

• The model so far only based on mandatory activities (HWH and HSH)
• Only 5% of population are simulated (1.1 million)
• 1 iteration takes more than 1 hour
• We use mode-choice model parameters estimated for Zurich,

Switzerland (Jakarta RP/SP survey is on-going)

Limitations

20

• Include mandatory and secondary activities based on our new survey

“MOBILITY JAKARTA” in the model

• Using choice model parameter of Greater Jakarta
• Simulate emerging transportation
• Online motorcycle taxi
• Online car taxi
• Urban Air Mobility
• Policy recommendations for the government

Future Work

21

Questions ?

www.ivt.ethz.ch www.MATSim.org

22

Appendix

23

Mobility Jakarta Survey

24

• RP Survey
• SP Survey

RP : Mode share by distance

25

RP : Mode share by activity

26

27

RP: Travel Time and Distance by Mode

RP: Travel Time and Mode by Activities

28

RP : Travel Time and Cost by Mode

29

RP : Trip Chain Distribution

30

Trip Chain Share of Trip Chain

h,w,h 39.12% h,e,h 22.70% h,w,l,w,h 13.66% h,s,h 5.83% h,w,l,h 4.40% h,l,h 3.53% h,er,h 1.97% h,s,w,h 1.32% h 0.74% h,e,l,h 0.74%

Trip Chain Share of Trip Chain

h,w,l,w,l,w,h 0.73% h,w,l,w,l,w,l,w,h 0.48% h,w,s,h 0.45% h,er,w,l,h 0.34% h,w,er,h 0.32% h,w,s,w,h 0.27% h,w,l,w,l,h 0.24% h,s,w,l,h 0.19% h,er,w,h 0.18% Others 2.77%

h = home e = education s = shopping w = working l = leisure er = errand

SP : Choices, Distance and Types of Respondent

31

Variable Value Sample (%) Choice PT Car MC Car taxi MC taxi Car ODT MC ODT UAM Walk 21.00 10.20 41.00 2.10 6.20 3.30 9.40 3.00 3.70 Distance band 0-1.5 km 1.5-5 km 5-15 km >15 km 11.00 16.00 35.90 37.20 Type of respondents Driver to Jakarta Driver out of Jakarta Non-driver in Jakarta Non-driver out of Jakarta 44.20 18.30 1050 27.00

32

SP: Distance by Mode

RP: Types of Respondents by Mode

33