Regulation , land - use mix , and urban performance . Part 2 : simulation

Part 1 of this two-part paper presented a spatial economic model of the urban development process which captures developers' profit-seeking behaviour, communities' welfare-seeking behaviour, and the mediating effects of alternative systems of land-use rights. Different systems of rights were shown to result in different land-use and density outcomes. In part 2 we describe the simulation model used to implement the theoretical model. The emphasis is on explaining the cellular automata methodology, but we also go on to illustrate the model output by comparing the structure and economic performance of two simulations. One simulates a free-market city in which developers have full property rights over land use. The other simulates a city in which the community has land-use rights and uses these to regulate development densities at socially optimal levels. 1 Introduction The main objective of this paper is to describe the approach taken to simulating our theoretical model within a cellular automaton. In this we try to give sufficient information to convey the mechanics of the method without confusing the reader with excessive detail. To avoid clutter we have left a full discussion of the quality and usefulness of the output of the model to other papers. Here we introduce the performance measures that it generates and illustrate these with reference to two simulations. In section 2 we comment on the general implementation strategy and in section 3 we discuss the local kernel evaluation function used to simulate the effect of same-use clustering on the developer's profit function. Section 4 is a description of the use of a spatial interaction field to simulate the community compensation function. In section 5 we define density, profit, and social cost equilibria in terms of the simulation model. In section 6 we present the Monte Carlo process which governs land conversion under multiuse competition and in section 7 discuss the problem of converting profit to development probability. Section 8 is a general discussion of performance indicators in cellular automata (CA) models. In section 9 we introduce two illustrative simulations—free market and planned—and in section 10 discuss their performance. We conclude with comments about the simulations, about the overall methodology, and about future work.