The idea of rule based self-organizing structure is further explored in the graduate thesis project, Future Autopia. The proposal investigates the performance and formal effect of the swarm intelligence in a non-uniform adaptive system. Self-organization is a set of dynamic mechanisms whereby global patterns or spatiotemporal structures emerge from the accumulation of local interaction between neighboring units. The existing freeway system will be reconstructed by numerous initially homogenous units and perform through the logic of self-organization to create a continuously differential territory and adaptable space, which produce various scales of hybrid infrastructure and urban functions according to flow density.

Abstract

Architecture concepts are expressed as generative and organizational rules, so that their evolution maybe accelerated and tested. The rules are described in a genetic language which produces a code-script of instructions for form generation and interaction. Computer models are used to simulate the development of forms which are then evaluated on the basis of their performance in Los Angeles’ future autopia.

Statement

The Beginning of Amoebae

What if architecture has no definite form? It is an amoebae deforming through time, adaptive to the environment. Forms are not things. They are emergent processes. Architecture is a collective effect of numerous parts. Parts have their individual identities. Under the strategies of juxtaposition or superimposition, discrete parts can be linked together to form a megastructure, which stays in the phrase of quantitative transition while parts still preserve individual identities. The qualitative transition from part to the whole can be achieved through the process of self-organization. The consequent emergent identities of the whole are not reversible and reducible.

Definition of Self-Organizational System

According to Francis Heylighen, self-organization is basically the spontaneous creation of a globally coherent pattern out of the local interaction between initially independent components. Most of the groups that have self-organizing capability are composed of numerous ‘parts’, which could be molecules, cells, individual organism or social units. The individual ‘parts’ will have impact on the ‘parts’ within certain distance to create local conditions. The accumulation of numerous ‘local conditions’ will influence further ‘parts’ in the system to achieve a global pattern or structure. In the whole process, there is no external top-down planning or forces imposing on the arrangement of parts, but the simple rules directing the action of parts. As a result, the global pattern or structure, either static or dynamic is the collective effects from numerous parts, which are so-called emergent properties.

Self Organization systems wildly exist in the scientific fields and human societies, like molecules arranging itself into a more ordered and fixed pattern in crystallization, chemical reactants forming swirling spiral, cell making up highly structured tissues, fish joining together in schools, cars moving on the freeway and social behaviors operating in urbanization. 

Generative Architecture

The practice of self-organizational model in architecture field as a form-generation tool can trace back to 1970s. John Frazer generated a reptile structural system for a gymnasium from a single seed through semi-automatic process at Cambridge University’s mathematical laboratory. His continuous pursuit in using genetic algorithm to create architectural form through computer established the theory of an evolutionary architecture, which was further developed into an emergent architecture in 1990s by Karl S. Chu. The limitation exposed in Frazer’s evolutionary models, which are always quantitative aggregation of individual parts through interconnection or adjacency has breakthrough in Chu’s emergent models, which are qualitative accumulation of single part through replication, mutation, recombination, and reproduction.

Most of the generative models have been used as dynamic structure systems, wall patterning systems and unit-based living systems. The open and uncertainty feature of self-organizational model enriches today’s dynamic environment. What’s the future potential of self-organizational generative model in a wider urban and architecture system becomes main concern in the thesis.

Self-Organizational Autopia

The thesis explores a self-organizational model in Los Angeles’s autopia in 2025. The model consists of two reciprocal systems, which provide an automatic control of traffic congestion. The external system, as the sum of running cars, will display periodic changes or accidental changes on the speed due to rush hour traffic or accidents at specific locations. The internal system, as a combination of construction components, will develop horizontally or vertically according to traffic flow and topographical conditions in the vicinity. The emergent global configuration of the internal system will create an intermediate space between the autopia and the city, providing drive-in shopping, dining, accommodation, and entertainment services. With the formation of the new “sidedrive” (derived from “sidewalk” of the city), the external system starts dissolving into the internal system, which will relieve the traffic congestion.  The sidedrive functions as sidewalk in the city, providing a unique space for interaction among individuals, facilitating the drive-in culture in future Los Angeles.