The Spatial Modelling and Dynamics Research Group focuses on the representation, understanding and optimization of traffic and transportation systems across multiple spatial and temporal scales. The group’s research aims include understanding and predicting the response of network-based transport systems to changes in demand and/or congestion levels, charging and management; understanding the characteristics and consequences of network structure and topology; development of state of the art traffic models; development and evaluation of effective methods for the monitoring, management and control of traffic and traffic emissions. Additionally, the group does research into the scientific methodologies adopted in transport modelling and the philosophical foundations that underpin them.
Contact: Dr Richard Connors
0113 343 1799;
R.D.Connors(at)its.leeds.ac.uk
Network modelling
The research focus in network modelling is to develop models that represent the spatial movement of people and goods in transport networks. The research aims include understanding and predicting the response of network-based traffic systems to changes in demand and/or congestion levels, charging and management, the development of state-of-the-art traffic network models, and the development and evaluation of effective methods for the monitoring, management and control of traffic. A number of crucial research directions in this area are identified below.
Network analysis and optimization
Development of methods spanning spatial and temporal scales e.g., short
term traffic control, medium term pricing of multi-modal personal and logistics systems, long-term policy optimization
of land-use/city structures. Casting transportation problems in a form appropriate for the application of optimization
methods e.g. to solve equilibrium problems, or modelling multi-level competition between actors (personal travellers,
freight operators, airline and train operating companies, airport/infrastructure owners, local and central government)
in the setting of a multi-objective problem with trade-offs between economic, environmental, energy, safety and
social goals.
Traffic dynamics
Dynamic Traffic Assignment (DTA) is concerned with the problem of assigning
time-dependent flows to the links of a traffic network that are mutually consistent with some underlying philosophy of
(a) time-dependent driver route choice and learning, and (b) the temporal and spatial propagation of traffic flows
through a network. Within ITS we pursue advances in both of these aspects through the following interests:
adaptive dynamic process models and their interaction with choice modelling, micro-simulation of private/public
transport & pedestrians, time-dependent interactions between activities/policy/transport use at the
local/national/international scale, statistical estimation of network models and potential of new data sources,
inter-modal networks (e.g. access modes and aviation networks, train/road networks).
Reliability and Uncertainty
Transportation networks are exposed to various sources of uncertainty in the
real world; demand fluctuates day-to-day as a result of travel decisions and the network supply can suffer short or
long term degradation due to external factors (from weather to variations in on-street parking). Travellers and
network planners faces uncertainty in the network’s level of service. This uncertainty has impacts ranging from
individual late arrival, loss of economic efficiency, or even threats to national security and safety. In order to
address these problems, one needs to understand the nature and effect of uncertainties related to network
operation, short and long term planning, the evaluation of network reliability, or travellers’ behaviours under
uncertain conditions. Research topics include: network design and traffic management for reliable transport
networks; the valuation of network reliability; data collection methodology for reliability analysis; recovery
strategies after emergency conditions.
Fundamentals of modelling
Transport researchers from different academic disciplines can find commonality in various cross cutting themes that highlight the need to advance methodology if we are to adequately represent important real world phenomena and develop applied research.
ITS has significant skills in many dimensions of transport modelling, including behavioural modelling, network modelling, and the modelling of externalities including the environment and accidents. Collaborations between these different areas clearly deserve attention. This applies for example to combining state-of-the-art techniques in network modelling and behavioural modelling to inform the development of the next generation of micro-simulators for road traffic. Similarly, incorporating behavioural modelling and environmental modelling provides the potential for ground breaking research in the relationship between driving behaviour and emissions.
Common modelling issues
A number of major issues arise across different areas of
transport modelling where bringing together the strengths from different backgrounds help us
address problems such as the representation of behavioural adaptation, habit formation, day-to-day
feedback, and interactions between agents; accounting for risk, heterogeneity in sensitivities and
perceptions, reliability, and errors in explanatory variables; deciding on the level of disaggregation
required for modelling and the level of aggregation required for appraisal; and examining and improving
the temporal, spatial, and social/societal transferability of model forms, results and parameter estimates.
Foundations & philosophy of modelling
Whilst individual scientific advances are pursued within each
modelling area, it is important to recognise that some of these activities can be categorised as truly
fundamental research that questions, and contributes to, the existing canon of transport modelling
methodologies. Here we determine how existing methodologies, theories and models relate to one
another, and how we can establish consistency in our fundamental analyses. Looking at how transport
modelling has evolved over the past 60 years and how it contributes to the benefit of society is vital in
order to continue building the theoretical framework that underpins our analyses, and to allow us to
deliver future cutting edge research in the strategic applied research areas. Here, we also look at how
we can ensure that our methodological work has the expected impact in real world applications.
Building the next generation of models
A major emphasis in transport modelling is on producing
forecasts, and the 21st century is already demanding that we address long-term uncertain futures
(climate change, carbon agenda, economic instability). Existing models were not always designed
for this purpose and are stretched to breaking point in applications that claim to encompass these
time-scales. Similarly, the actual processes modelled are often more complex than the methodological
frameworks permit. Rather than simply further developing existing structures, part of the emphasis
is now on developing new methods that are more suitable for solving the new set of problems. With
models ranging from microscopic, meso-scopic to macroscopic, and from individuals to the whole
population, there is a need for a hierarchical or multi-level modelling architecture that allows for the
integration of these models.
For further information about the Modelling Research Group and its activities, please contact Dr Richard Connors
