The Dynamic Urban Model Transport and Urban Development John - - PowerPoint PPT Presentation
The Dynamic Urban Model Transport and Urban Development John Swanson Steer Davies Gleave February 2004 Transport and the Economy What is the contribution of transport to a local or regional economy? What is its role in regeneration?
The Dynamic Urban Model
Transport and Urban Development John Swanson Steer Davies Gleave February 2004
Transport and the Economy
What is its role in regeneration?
What is the effect of specific schemes? Of area policies?
What is the effect of Road Pricing?
The two-way road effect.
What is the contribution of transport to a local or regional economy? This has been widely researched – but the results are ambivalent.
Conventional Transport & Land Use Models
Traditional transport models are almost exclusively equilibrium models
which makes them technically very difficult.
Classic four stage model
Trip generation Distribution Mode choice Route assignment
Then add land use effects …
The Dynamic Urban Model
Draws on Forrester’s original Urban Dynamics work
But this said nothing about transport
Combines it with ideas from conventional transport modelling
The logit model for choices Network modelling software Generalised cost, utility etc
Transport and the Economy
Access of employers to a suitable workforce
Access of the workforce to suitable jobs
Access of businesses to suppliers, markets and other businesses
Why should transport contribute? It must be because it affects ‘accessibility’
Access and Attractiveness as Business Location
Transport infrastructure travel times Workforce Businesses
workforce accessible businesses attractiveness as a business location +
+ + Net business startup rate +
Access and Attractiveness as a Place to Live
Businesses Workforce travel times accessible jobs Net migration rate attractiveness as place to live
+ + jobs per business +
Competition for Land and Premises
Workforce Businesses Housing stock Business premises Land available for development housing availability
attractiveness as place to live + developers expectations for housing demand net migration rate net business startup rate availability of business premises attractiveness as a business location developers expectations for business units +
new houses + + +
business premises
+
+ +
Access: The Congestion Feedback
transport infrastructure travel times transport activity workforce businesses
+ +
Putting it together
accessible jobs attractiveness as place to live housing availability net inward migration Skilled workforce accessible workforce transport activity travel times Housing stock housing construction & enhancement rate Land available for development land-use policy attractiveness as business location net business startups + + + + +
+ + + + + Businesses in area + business unit construction rate Business units + + availability
+
+ + + workforce elsewhere
jobs in other locations + Transport infrastructure
markets and supply chain +
+ transport investment + +
Transport
Transport is spatial
So the model must divide the urban area into zones
Network effects
We must represent movements through networks
Congestion at a location is affected by and affects activity elsewhere
Transport offers choices of route, mode, whether or not to travel etc
So we need to represent these choices
Preferences and habits vary
We need to recognise different preferences among people & trip purposes
Not all modes are available to everyone
Car ownership varies and is growing
Availability of licences, household size etc.
Transport Networks: Roads
Its length (km)
The free-flow speed;
The capacity (vehicles/hour) beyond which speeds will drop;
Any additional fixed costs (eg road pricing).
The road network is described in terms of links. Routes between each pair of zones are specified as sequences of links. Several routes may be available – the model is capable of modelling route choice dynamically as conditions change. Each link has characterising properties: When flows exceed capacity, speed-flow curves are used to reduce speeds.
Transport Networks: Public transport
Its length (km)
The free-flow speed;
The service headway;
The capacity of each vehicle;
Access and egress times to/from the network;
One-way fares;
Interchanges;
Additional mode penalties (eg to distinguish between modes)
The PT network is also described in terms of links, and routes between each pair of zones are specified as sequences of links. Each link has a number of properties:
Modelling choices: the Logit Model
Logit models are widely used in transport to model choices. They relate the ‘utility’ of each available alternative to the probability
j j i
It’s up to us how to define utility – but usually it’s a linear additive function of the attributes – cost, time etc. Much existing research exists on suitable parameter values to use.
Travel choices
Begin Too far In range Car Public transport Walk Route1 Route 2 Route 3 PT Route 1 PT Route 2 PT Route 3
The hierarchy of choices available when considering travel between two zones:
Whether to travel at all
Choice of mode
Choice of route
The Deterrence Curve
0.2 0.4 0.6 0.8 1 10 20 30 40 50 60 Drive time (minutes)
The whole choice structure is handled by a hierarchical logit
travel at all. This gives rise to a deterrence function, relating the cost of travel to the proportion of people will to accept the cost. For instance, from zone i: Accessible workforce =
j j
The logit model
The logit model is instantaneous, assuming:
Everyone has perfect knowledge of the alternatives; and
They make their choices immediately
But it’s well known neither of these is correct
A Dynamic Logit Model: Choice of Route
routes current link times actual route share total trips on each link route times change in share Time to adapt to network changes target route shares capacity base link times Car route generalised times congestion function capacity ratio Capacity time multiplier Averaging period for work trips Trips to work per hour Car route utility scaler beta Car route utilities <Car travel to work trips>
Business Clusters
accessible businesses attractiveness as business location transport times Businesses
+ + + +
close to businesses and services, versus the congestion that results.
Applications of the Model
Applications of the model
Hastings and Bexhill
Highly controversial by-pass proposal
Merseyside
Light rail schemes
North East England
Area in economic decline – wish to demonstrate transport investment can kick- start the economy
Applications of the Model
Location Workforce Jobs Zones Transport Hastings & Bexhill H: 36,000 B: 12,000 H: 33,000 B: 13,000 17 GC from transport model Merseyside 490,000 448,000 140 GC from transport model NE England 1,130,000 923,000 81 Full network Hypothetical 26,000 25,000 7 Full network
Note: GC means Generalised Cost
Typical Outputs
Year: 2011 Base Land Release Change % Change Jobs 893 1,094 201 22% Vacancies 58 82 24 42% Households 952 1,007 55 6% Workforce 1,115 1,184 69 6% Job seekers 197 89
55% Car kms 5,495 6,176 681 12% Average speed (km/h) 42.8 39.2
Thousands
Typical Outputs (2)
Year: 2011 Base Land Release Change % Change Trips per day Car 606 722 116 19% PT 129 155 26 20% Walk 101 135 34 34% Mode shares Car 72% 71%
PT 15% 15% Walk 12% 13% 1
Typical Outputs: Network Plots
Typical Outputs: Analysis of Movements
Some Findings: the Effects of Reduced Transport Costs
Commute distances rise quite rapidly
Eg road improvements increase car mode share and average commute distance – may conflict with environmental objectives;
Effect of specific schemes on job numbers can be small
Eg +1,000 jobs in NE England from large package of measures
Congestion tends to re-locate
Other constraints may be more important
Especially land – but it must be accessible
Was transport a constraint in the first place?
Eg in Merseyside the workforce > jobs, so recruitment was not really a constraint. New transport re-distributed recruitment patterns but did not increase jobs – but did increase employment in poor areas
Some Findings: the Effects of Road Pricing
Tests of road pricing showed initial change in behaviour but people shifting travel-to-work patterns and modes to revert to car use
But also a tendency to push economic activity outwards – increased sprawl?
Retail is doubly hit - staff and customers
Change of use in the charge zone as retail leaves
Some client responses & issues
The issue is not transport within the region, but connections to the rest of the country/ Europe/ rest of the world
How did you calibrate that?
Existing parameters
Interviews (including SP/Conjoint)
Calibration using data (eg the deterrence curve)
Judgement
Some confusion about site specific schemes and area policies
Site schemes may enable a few new jobs locally, but will not regenerate the area
And the answer is ..
Depends on local circumstances and configuration
Effects are probably small
In a developed area transport is unlikely to be a major regenerating force unless it can be shown it is a constraint on activity
Which is unlikely in a regeneration area
Work continues on the model:
Technical improvements; but also (and harder)
Convincing potential users