Professor Mike Batty | QUANT: Scaling Up Urban Models To Test Large Scale Planning Decisions

As computers have got ever more powerful and as software and data can now be mainly accessed across the net from remote locations, the constraints on building models of urban systems that were present a generation or more ago, are rapidly changing. In the 1960s, our models were highly constrained with respect to the level of spatial and topical detail and few models were built with more than a couple of hundred zones. The problem of knowing were to draw the spatial boundary was critical and many models simply defined away the interactions between the systems and its environment due to limits on how big the model might be. We are now in a position to build extremely large models with respect to spatial zoning, we can run these rapidly making the results known in real time using powerful visualization techniques, and we can make these models available across the web for a variety of users. Moreover since the first urban models were constructed in the 1960s, cities have grown dramatically in size and the impacts of new development within them often span the nation or continent in which they are located. In this paper, I will describe the effort by myself, Richard Milton and Roberto Murcio to build an urban model named QUANT for Great Britain (England, Scotland, Wales) for some 8436 zones where spatial interactions are routinely in the order of 70 million pairwise links. A prototype is available on the web at http://quant.casa.ucl.ac.uk/ and the model is set up to test large-scale scenarios pertaining to population, economic development, and new transportation schemes particularly in Greater London and across the nation based on high-speed trains. The model runs in a matter of seconds and successive ‘what if ‘ scenarios can be developed by any user from any place where there is an internet connection. The model is pretty crude at present but our current work the project is to improve the interface massively. Here, we will describe its construction, operation and use to test new urban infrastructure proposals speculating on how we can extend the model to deal with a multilevel spatial system.