Background material on the following topics is available on this site:
|
 |
Land-use and transport models are perhaps the most "operational" class of urban simulation. They are used as decision support systems in urban planning, public policy formation, and urban management.
Why do we need land-use and transport models?
There are some powerful rationales for applying simulation models to the study and management of urban systems. In Europe, concerns about the sustainability of our cities is driving a concerted effort to model their functioning in a bid to forecast future urban patterns. Meanwhile, in the United States there exists legislation that both directly and indirectly encourages the development of simulation models of various urban phenomena to assist cities in complying with air quality standards. This legislation includes the Clear Air Act Amendments: CAAA (1990), the Intermodal Surface Transportation Efficiency Act: ISTEA (1991), and ISTEA's successor, the Transportation Equity Act for the Twenty First Century: TEA-21 (1997). However, other initiatives, such as the Travel Model Improvement Program: TMIP (1992), which was established by the Federal Highway Administration; the Federal Transit Administration; the Office of the Secretary, U.S. Department of Transportation; and the U.S. Environmental Protection Agency, have been introduced specifically to encourage improvements in land-use and transportation modeling.
Other justifications for urban simulation models include the functionality that they offer by allowing us to test theories and practices about urban systems in a controlled computer environment. Proceeding from a simulation model, we can evaluate the merits of theories relating to urban phenomena and test the application of policy measures (such as growth management, congestion pricing, and pollution mitigation schemes) to various scenarios for urban futures.
|
|
Overview
Land-use and transportation models belong to the mathematical family of models. They are composed of independent land-use and travel models, with mechanisms for coupling the two--either loosely or in a more integrated fashion. Land-use models are used to simulate demographic and economic transition in land-based activities. These measures describe the population (usually in terms of income and employment) and built-space environment (e.g., floor space) for a given urban area. Travel models (specifically, travel demand models) are used to simulate travel patterns on a transportation network. This class of models aims to simulate travel patterns as a function of human activities (commonly considered in terms of land-uses) as well as the characteristics of the transport network (commonly considered in terms of accessibility). Integrated land-use and transportation models are used to simulate the interaction of the land-use system and the transport system. Generally, this interaction is simulated by means of feedback mechanisms.
There are many 'styles' of land-use and transport model, including those based on spatial interaction, spatial choice, and activity-based models. For a more thorough look at how land-use and transport models are designed and built, please have a look at my working paper, entitled, 'How land-use and transportation models work' (.pdf; 497KB).
|
|
General structure of the land-use and transport model
As intuition would suggest, land-use and transport models couple two distinct systems: land-use and transport. Embedded beneath the umbrella of these two systems, however, lies an inter-connected web of sub-models representing various sub-systems and processes at work within the city. Depending on the peculiarities of the model in question, these sub-models may exist in isolation from each other, they may be loosely associated, or may be well connected via such mechanisms as feedback loops.
Generally, a number of key components underpinning the land-use-transportation model may be described. These include, at the top-most level, a mechanism handling land-use and a separate model to describe transport. The land-use module depends, in varying degrees, on sub-models for location, land development, and an equilibrium mechanism that balances forces of demand and supply. The transport system is traditionally simulated via a four-step process beginning with potential demand modeling and trip generation, proceeding through trip distribution and modal split, and concluding with trip assignment.
The conceptual framework for a traditional land-use and transport model.
|
|
Advancing the state-of-the-art
In modeling urban systems such as land-use and transport, we are faced with a dilemma. The intellectual apparatus with which we model urban systems evolved in a time when the city was very different from contemporary manifestations. Single-centre cities built with raw materials, labor, and trade have given way to polycentric cities restructured by automobiles, services, and information technology. It is clear that many of the tools with which we study the city today are deficient in their ability to fully simulate and describe the changing character of urban areas. Consequently, there is a need for models that are as flexible and dynamic in their simulation capabilities as is the city in its ability to evolve.
Many land-use and transport models operate on a weak theoretical foundation. In many cases they depart from what we know about the way in which urban systems evolve and the dynamic forces that shape them. In short, there is much room for improvement. One feasible remedy to the these problems is to weave ideas from complexity theory with existing techniques to arrive at a hybrid, modular simulation strategy for modeling urban systems. Such an approach would build upon those areas of traditional land-use and transport models that work well (particularly at the macro- and the meso-levels), but would delegate the micro-scale dynamic simulation to sub-models derived from complexity theory: cellular automata and multi-agent systems.
The agent-based approach to simulation seeks to represent individual actors (or groups) in a given system. Agents may interact with each other and/or with an environment. From these interactions, macro-scale behaviors emerge in the aggregate. Agent-based models have been used to simulate insect behaviour, search the Internet, and to manipulate financial data. Agent-based approaches have also been used to simulate urban systems, including traffic dynamics, pedestrian movement, and lines of sight. Equally, we might envisage agent-based models that represent the agents that compose the land-use and transportation systems--migrating households, firms, or individuals; socio-economic groups; commuters; pedestrians; developers; etc.
|
