Search

The paper presents a methodology for describing in generative terms the structure of urban fabrics: the objective is to transfer conceptually the knowledge about the domain of urban space into a hierarchical and interrelated semantic structure with relevant concepts, elements and their mutual relationships, providing explicit and unambiguous definitions. The conceptual and operational instrument adopted for this purpose is the ontology, a method of knowledge representation and management coming from the Artificial Intelligence. This approach aims to create a customisable digital design tool, to support the designer in the early stages of urban design process, such as street pattern and massing definition, by generating in real time a number of design scenarios, starting from a large number of constraints and requests. This paper focuses on the knowledge formalisation aspects of the research that is the basis for the generative modelling of urban space. [ABSTRACT FROM AUTHOR]

Cellular automata (CA) are an effective means of urban-growth simulation. Moreover, the development of a planning support system associated with the theory of artificial intelligence has recently become a new global challenge. In this paper we investigate a geosimulation model using CA based on the geographic information system (GIS) for visualising land-use patterns in urban partitions. Most recent simulation models using CA have been developed in regular grid nets or networks, where adjacent grids or networks in urban space are easily addressed. Our project explores how to utilise an irregular polygon dataset in a GIS database in order to deal with spatial microsimulation. A new concept used in this project takes into account the effects of urban planning, such as land-use zoning and street networks, at the level of parcels and blocks in urban space. Adjacent irregular parcels on both sides of a street are intended targets in the simulation. In this paper, a method of simulating land-use patterns in urban space after the implementation of land-readjustment projects is discussed. [ABSTRACT FROM AUTHOR]

This paper develops and applies neural network (NN) models to forecast regional employment patterns in Germany. Computer-aided optimization tools that imitate natural biological evolution to find the solution that best fits the given case (namely, genetic algorithms, GAs) are also used to detect the best NN structure. GA techniques are compared with more `traditional' techniques which require the supervision of experienced analysts. We test the performance of these techniques on a panel of 439 districts in West and East Germany. Since the West and East datasets have different time spans, the models are estimated separately for West and East Germany. The results show that the West and East NN models perform with different degrees of precision, mainly because of the different time spans of the two datasets. Automatic techniques for the choice of the NN architecture do not seem to outperform selection procedures based on the supervision of expert analysts. [ABSTRACT FROM AUTHOR]