Geoscience Reference
In-Depth Information
In the next stage of the process (
Step 5
), the most complex piece of modelling technology is intro-
duced in the form of a dynamic simulation of the demographics of the city over the time horizon
of the projections. In the exemplar described here, up to 15 years has been specified, although this
model is now generically available for British cities and regions on a time frame of up to 30 years.
The model components are described elsewhere (Wu et al., 2010). Longer timescales are now also
under consideration as part of a project to model future demand for infrastructures, including but
not limited to transport, albeit at a somewhat more aggregated geographical scale (local authority
districts with typical populations in the order of 100,000 rather than much smaller census wards)
(Birkin and Malleson, 2012).
At this point, two significant features of the workflow process are noteworthy. First, the con-
cept provides a simple structure in which the baseline model actions (i.e. estimate demand, allo-
cate demand, generate indicators, map indicators) can be executed repeatedly on any subsequent
model projection. These projections can be perturbed if necessary, for example, to assume a steady
increase in the rate of trip generation. Second, the workflow components are structured hierarchi-
cally. Thus, within the simulation workflow as it has been described here, individual components
including the PRM, the dynamic simulation model and the MapTube visualisation engine are all
workflows in their own right.
Once simulation results from years 0, 5, 10 and 15 have been generated, they can be
published
to an area of the portal (
Step 6
) so that planners or members of the public can see how much the
local transport system could deteriorate without further intervention or behavioural change. In
an ideal world, members of the general public might even be invited to view these simulations
(
Step 7
), and some work has been undertaken on the specification of a virtual exhibition space in
which these results could be explored and displayed (Batty and Hudson-Smith, 2007). Perhaps the
local residents could be encouraged to experiment with solutions, for example, would it be a good
idea to leave their cars at home and get the bus to work every other Thursday?! At the moment,
the public interface to NeISS is not fully enabled. However, substantial progress has been made
with a tool that does allow the world at large to express their views about future transport poli-
cies, simulation outputs or any of a variety of social and political questions and items of general
interest. The components of the NeISS infrastructure and their linkage to solve a practical prob-
lem have been illustrated in a demonstrator for the Greater Manchester congestion charge, as
discussed in Section 10.9.
Finally (
Step 8
), the workflow architecture permits the straightforward re-enactment of entire
sets of simulations that may be regenerated under alternative scenarios relating to external pol-
icy shifts or behavioural change. For example, what happens with reduced traffic demand, for
example, assuming that x% of the population is willing to leave their cars at home on y days per
month?
For the purpose of the demonstrator, we modified a transport component to the simulation using
a customised version of a generalised SIM which has also been applied to migration flows (in the
demographic models) and to retail flows in the NeISS. We defined bespoke indicators for illumi-
nation of the simulation outputs - things like congestion (average travel times), pollution, jour-
ney types (modal split) and road accidents. Examples are shown in Figure 10.4 using model-based
indicators for the base year (2001), projection year (2011) and scenarios (2021). Figure 10.4a and
b demonstrate the spatial flexibility of the models, showing pollution estimates for three different
cities. Figure 10.4d through f show a variety of performance indicators for each city across the time
horizons of the simulation.
10.6 NEW COMPUTATIONAL DIMENSIONS
A vital factor for e-Research, which up to this point has been largely absent from our discussion,
is that of the hardware which is ultimately used to perform the necessary calculations and manage
the required data. In essence, the vision for e-Research is one where computing power and storage
