Evaluation of Shape Grammar Rules for Urban Transport Network Design

Abstract

Shape grammar rules are increasingly applied in urban simulation. Even though many network design standards propose shape grammar rules, little is known of the measurable impact of these rules on the performance of transport networks. This paper provides a general definition of shape grammar rules for transport network design. Different rules are evaluated regarding a comprehensive objective function. Networks are designed and simulated on featureless planes to avoid a bias due to history. Findings are compared with real-world case studies. Different network characteristics are evaluated in this paper. The densities of network loops are high in all generated networks, and comparable with real-world grids and medieval fabrics. The average length of network loops decreases as an inverse function of road density, which is in line with graph theory. Intersection density is proportional to the network length. The average number or arms of an intersection depends on road density. A denser network has a disproportionately higher density of 4 arm intersections, compared to less denser networks. Additionally, different road types are assigned to each road segment. Hierarchical road type distribution has a significant but low influence on network user costs. Terrain boundaries, as well as predefined roads (e.g. boulevards) increase average user costs. However, the average increase strongly depends on the number of bridges and on the boulevard capacity. The results show that shape grammar rules for transport network design can be evaluated to increase the understanding of their impacts, which supports future design standards.