人工演进的元城市系统:城市空间形态的一种智能生成

Meta Urban System of Artificial Evolution: A Kind of Intelligent Generation of Urban Morphology

为探究城市空间形态的生成如何更能反映城市作为复杂巨系统的本质,从理论上初步探讨复杂系统的可持续建构、多专业优化协同、多尺度网络建构等理念,探索城市空间形态生成的内在逻辑,提出自我优化迭代的城市空间形态生成的算法演进模式。借鉴复杂系统和网络科学的理论成果,提出基于图网络的空间形态生成、运筹优化下的实体形态生成和参数化的系统生成的方法体系,建构包含空间与行为形态网络分析、用地功能与公共服务设施网络分析、三维形态生成、专业协同评估、人机互动调整等基本功能的系统。最后介绍该系统在深圳前海妈湾片区规划设计工作中的应用,以及深圳大学和清华大学的工作营活动中的实践尝试,对未来城市发展的价值动力的探索给予一定的启示。

To investigate how the way of generating urban spatial morphology reflects the essence of the cities as mega systems, this paper, from the perspective of theories, seeks to discuss the ideas on sustainable construction of complex systems, collaborative optimization of different disciplines, and multi-scaled network formation, explores the basic logics of generating urban spatial forms, and makes the initial proposition of establishing the pattern of the evolutional algorithm for producing urban spatial forms via the process of self-optimizing iteration. The paper draws on the theoretical achievements of complex systems and network science, and proposes the methodology which consists of the graph-based generation of spatial forms, generation of building forms via operations research, and parametric-aided generation of systems. In this way, a platform emerges with basic functions such as analysis of relationships between space and human behaviours, network analysis of land use and public facilities, 3D shape generation, multi-disciplinary assessment, and design adjustment via human-computer interaction. Finally, this paper introduces the practice in the planning and design of Mawan in Shenzhen, as well as the workshops at Shenzhen University and Tsinghua University, which gives some insights to the exploration of the driving force of future urban development.