transmob: an agent-based simulation of transport demand and residential mobility in south east...

27
TransMob An Agent-based Simulation of Transport Demand and Residential Mobility in South East Sydney Pascal Perez Nam Huynh Vu Lam Cao Rohan Wickramasuriya Matthew Berryman September 2014

Upload: smart-infrastructure-facility

Post on 25-May-2015

165 views

Category:

Government & Nonprofit


0 download

DESCRIPTION

A presentation by SMART Infrastructure Facility Research Director Dr Pascal Perez to the Social Simulation Conference, Barcelona, September 2014.

TRANSCRIPT

Page 1: TransMob: An Agent-based Simulation of Transport Demand and Residential Mobility in South East Sydney

TransMob An Agent-based Simulation of

Transport Demand and Residential Mobility in South East Sydney

Pascal PerezNam Huynh Vu Lam Cao Rohan WickramasuriyaMatthew Berryman

September 2014

Page 2: TransMob: An Agent-based Simulation of Transport Demand and Residential Mobility in South East Sydney

WHAT IS IT?

Page 3: TransMob: An Agent-based Simulation of Transport Demand and Residential Mobility in South East Sydney

TransMob – fast factsTransMob includes:

110,000 Agents85,000 Private Motorized Vehicles50,000 Households (17 categories)

60 Travel Zones32 Bus Routes

8 Transport Modes6 Transport Purposes4 Dwelling Categories

TransMob generates:

~ 450,000 Trips/Weekday~ 9,000 Transport Mode Choices/Year

~ 7,000 Residential Changes/Year

(Baseline scenario, indicative values only)

Period: 2006-2026Calibration: Census 2006Validation: Census 2011

Page 4: TransMob: An Agent-based Simulation of Transport Demand and Residential Mobility in South East Sydney

TransMob – User Interface

Study Area Travel Zone Road Network

YES

NO

Page 5: TransMob: An Agent-based Simulation of Transport Demand and Residential Mobility in South East Sydney

WHY IS IT IMPORTANT?

Page 6: TransMob: An Agent-based Simulation of Transport Demand and Residential Mobility in South East Sydney
Page 7: TransMob: An Agent-based Simulation of Transport Demand and Residential Mobility in South East Sydney
Page 8: TransMob: An Agent-based Simulation of Transport Demand and Residential Mobility in South East Sydney

HOW DOES IT WORK?

Page 9: TransMob: An Agent-based Simulation of Transport Demand and Residential Mobility in South East Sydney

Model Components

Synthetic Population Liveability

Travel DiariesTransport Micro simulator

Agent Decisions

Transport Mode Choice

Residential Location Choice

Out

put D

ata

War

ehou

se

(sce

nario

s an

d ru

ns)

Soci

o-de

mog

raph

ic d

ata

Tran

sim

s da

ta

Visu

alis

ation

Das

hboa

rdSc

enar

ios

and

runs

Stud

y Ar

eaTr

avel

Zon

esM

ajor

Roa

d Li

nks

3D T

raffi

c Si

mul

ator

Traffi

c de

nsity

-bas

edIn

tera

ctive

ses

sion

sTe

mpo

raril

y st

aged

TransMob - Architecture

JtW 2006

HTS 2006-2010

Census 2006

Liveability 2011

HILDA 2001-2010

Page 10: TransMob: An Agent-based Simulation of Transport Demand and Residential Mobility in South East Sydney

Perceived Liveability - Calibration

Satisfaction with

transport

Influence of

Age Group

Page 11: TransMob: An Agent-based Simulation of Transport Demand and Residential Mobility in South East Sydney
Page 12: TransMob: An Agent-based Simulation of Transport Demand and Residential Mobility in South East Sydney

42%(51%)

Traffic Micro-Simulator - Validation

26%(29%)

(HTS 2006-2010)

Page 13: TransMob: An Agent-based Simulation of Transport Demand and Residential Mobility in South East Sydney

Traffic Micro-Simulator - Validation

Trip Counts per Mode Trip Counts per Purpose

Page 14: TransMob: An Agent-based Simulation of Transport Demand and Residential Mobility in South East Sydney

Synthetic Population - Validation

Page 15: TransMob: An Agent-based Simulation of Transport Demand and Residential Mobility in South East Sydney

Validation – Residential Mobility

Page 16: TransMob: An Agent-based Simulation of Transport Demand and Residential Mobility in South East Sydney

Exploring What-If? scenarios

Scenario 1: Baseline (calibrated on 2006 Census)

Scenario 2: Double Income No Kids (DINKs)

Scenario 3: Virtual Light Rail (VLR)

These 3 scenarios are purely hypothetical and do not reflect upon results being currently assessed by Transport for NSW

Page 17: TransMob: An Agent-based Simulation of Transport Demand and Residential Mobility in South East Sydney

Scenario 1 – Study Area - Year 2016

4% population increase50% car drivers12% public transport39,000 peak tripsLiveability: stable

Page 18: TransMob: An Agent-based Simulation of Transport Demand and Residential Mobility in South East Sydney

Scenario 2 – Study Area - Year 2016

4% population decrease46% car drivers16% public transport36,000 peak tripsLiveability: improved

Page 19: TransMob: An Agent-based Simulation of Transport Demand and Residential Mobility in South East Sydney

Scenario 3 – Study Area - Year 2016

7% population increase46% car drivers18% public transport45,000 peak tripsLiveability: stable

Page 20: TransMob: An Agent-based Simulation of Transport Demand and Residential Mobility in South East Sydney

Scenario 1 – Travel Zone - Year 2016

20% couple#053% 1-2 bedroom12% public transport0.40 satisfaction couple#0

Page 21: TransMob: An Agent-based Simulation of Transport Demand and Residential Mobility in South East Sydney

Scenario 2 – Travel Zone - Year 2016

48% couple#058% 1-2 bedroom15% public transport0.47 satisfaction couple#0

Page 22: TransMob: An Agent-based Simulation of Transport Demand and Residential Mobility in South East Sydney

Scenario 3 – Travel Zone - Year 2016

21% couple#057% 1-2 bedroom33% public transport0.41 satisfaction couple#0

Page 23: TransMob: An Agent-based Simulation of Transport Demand and Residential Mobility in South East Sydney

Scenario 1 – Road Traffic Profile - Year 2016

Northbound peak traffic (morning) 2,000 vehicle/hourSouthbound peak traffic (morning) 1,250 vehicle/hour

Page 24: TransMob: An Agent-based Simulation of Transport Demand and Residential Mobility in South East Sydney

Scenario 2 – Road Traffic Profile - Year 2016

Northbound peak traffic (morning) 1,650 vehicle/hourSouthbound peak traffic (morning) 1,100 vehicle/hour

Page 25: TransMob: An Agent-based Simulation of Transport Demand and Residential Mobility in South East Sydney

Scenario 3 – Road Traffic Profile - Year 2016

Northbound peak traffic (morning) 2,200 vehicle/hourSouthbound peak traffic (morning) 1,600 vehicle/hour

Page 26: TransMob: An Agent-based Simulation of Transport Demand and Residential Mobility in South East Sydney

TransMob – Current Limitations

No Macro-Economic Model for EmploymentModel uses a simple ‘job event’ probability to trigger a salary update and a residential relocation test. This probability is estimated using employment statistics from the Australian Bureau of Statistics.

No Proactive Real Estate MarketBaseline scenario uses a bootstrapping approach based on the evolution of dwelling stocks and prices in each travel zone between 2006 and 2011. Alternative scenarios allow only for demand-driven allocation.

No Urban Zoning or Planning MechanismsDespite repeated requests, we were not able to access and implement simple rules for local urban planning and zoning.

No Freight Transport Being ModelledThis is a major impediment as South East Sydney is connected to Mascot Airport / Port Botany Precinct!

Page 27: TransMob: An Agent-based Simulation of Transport Demand and Residential Mobility in South East Sydney

Pascal PerezResearch DirectorSMART Infrastructure Facility+61 2 4252 [email protected]