Urban Site Adaptive Layout Generation: User-Configurable Algorithms Based on MQP

Citation
Wang Y, Hu Q, Tang P, et al. URBAN SITE ADAPTIVE LAYOUT GENERATION:USER- CONFIGURABLE ALGORITHMS BASED ON MQP[C]//CAADRIA proceedings. Singapore, 2024. In Proceedings of the 29th CAADRIA Conference - Volume 2, Singapore, 20 April - 25 April 2024, pages 505–514

Sources
PDF


Abstract

The growing reliance on digital methods presents a promising avenue for addressing challenges related to maximizing design quality according to user preference. This paper introduces a suite of integrated algorithms based on Mixed Quadratic Programming (MQP) to enhance the adaptability of the existing model for generative layout challenges in urban design. The algorithms for site handling are divided into three key directions: (1) Boundary Adaptive, (2) Interaction with Existing Elements, (3) Flexible Road Network. These directions are implemented through a combination of hard constraints and soft constraints. The resulting layout model, incorporating these algorithms, is user-configurable to meet specific project requirements. To validate the effectiveness of this approach, we applied the algorithms in a campus planning project. The results demonstrate remarkable efficiency and precision in generating layouts that include plots and design elements. These findings affirm the robust applicability of our algorithms for creating urban-scale site layouts.

Keywords

Generative urban design, Quadratic Programming, Layout Modelling, Site Adaptation, User-configurable