Shape Grammars for Hierarchical Transport Network Design

Abstract

Urban systems are growing fast in many countries today, and therefore face the task to build efficient transport networks. Significant productivity gains are due to infrastructure investments in urban systems. A shape grammar based approach is proposed to contribute to more efficient networks by reducing the enormous and complex search space in the design of a new transport network. This paper sheds light on impacts of different shape grammars in network design. The proposed approach applies a network generation algorithm based on the integration of ant colony optimization (ACO) and a genetic algorithm (GA). For network design applications, the proposed algorithm is able to overcome restrictions of both ACO and GA, e.g. in network size and computational time, and is concurrently computationally fast. The algorithm generates best transport network layouts given a defined objective function, including demand weighted travel times. A budget constraint provides an upper bound for infrastructure investment. Shape grammars for hierarchical network design are respected during the generation of the network layouts. Different initial network layouts are tested. It can be shown that shape grammars can affect the resulting transport networks. Future research is proposed, including additional shape grammars, variable demand and growth processes, to verify and complete the results gained.