Note: This text is a work in progress, and will be further developed and changed as research continues within the ASL.
All Rights Reserved.
All Rights Reserved.
On Adaptive Design
Dino Rossi, 2011
nothing is static...
As the field of architecture increasingly integrates computer science into its methodologies, attention needs to be paid to the use of language. Architects frequently utilize concepts and terms from other fields, often lacking an in-depth understanding of those concepts and terms. This is due in part to the nature of architecture. Architecture, while primarily consumed by the conceptualization and realization of the built environment (landscape, building, urban form), necessitates a broad understanding of other fields (sociology, urbanism, environmental science, etc.) in order to address the surrounding requirements of any project. Clearly it is not possible for every architect to become a specialist in every field that architecture must consider. On the other hand, architects frequently assume a position of authority in the use of terms and concepts drawn from other fields. It is my belief that this is inherently problematic, and problematic on two specific levels. The first level is the weakening of the architectural project through an unclear understanding of its own description (in other words, a weak description of the properties of the project). The second is the dilution of the original intention/meaning of the terms and concepts used in regard to their fields of origin.
The term “adaptive” is frequently used in todays architecture scene. All of a sudden everything seems to be adaptive. While the idea of an adaptive architecture is quite intriguing, most of the projects to which the term is applied do not possess adaptive properties, rather they are interactive, kinetic, responsive, or to use a less sexy term, reactive. The properties which are deemed adaptive tend to be pre-programmed responses which react to particular stimuli.
So, what would constitute a truly adaptive architecture? To answer this question I would suggest to start by taking a closer look at the term “adaptation” and it’s uses in various fields.
The etymology of the word adapt stems from the Latin adaptō. From ad (“to, towards, at”) + aptō (“adjust, adapt; prepare”). (http://en.wiktionary.org/wiki/adapto#Latin) In it’s simplest definition the term adapt means to “adjust toward.”
The concept of adaptation as we presently understand it is based in the concept of evolution and dates back to Charles Darwin’s “On the Origin of Species” and earlier (reference tk).
“Adaptation is the evolutionary process whereby a population becomes better suited to its habitat. This process takes place over many generations, and is one of the basic phenomena of biology.” (http://en.wikipedia.org/wiki/Adaptation)
This definition illuminates a change over time. Not within the lifetime of an individual, but through generations which evolve in relation to their environment.
“The term ‘adaptation’ in computer science refers to a process, in which an interactive system (adaptive system) adapts its behaviour to individual users based on information acquired about its user(s) and its environment.” (http://en.wikipedia.org/wiki/Adaptation_(computer_science))
This definition offers a direct translation into architectural conceptualization. We can consider the building to be a system which adapts its behavior to information acquired about its users. Information external to the building (system) could also be integrated into the process, for example weather data, energy prices, demands of neighboring buildings, etc.
“An adaptive system is a set of interacting or interdependent entities, real or abstract, forming an integrated whole that together are able to respond to environmental changes or changes in the interacting parts. Feedback loops represent a key feature of adaptive systems, allowing the response to changes; examples of adaptive systems include: natural ecosystems, individual organisms, human communities, human organizations, and human families.” (http://en.wikipedia.org/wiki/Adaptive_system)
From this definition we can view the building along with its inhabitants as a system with an overall performance which is dependent on the independent actions of the various agents at work within the system (i.e. building occupants, building systems, etc.). This broadens the definition from an adaptive architecture which responds to internal and external forces, to an adaptive architecture which conceptualizes the system as including both building systems and inhabitants as agents. In this case the performance of the system is defined by the actions and interactions of all agents in the system.
“Adaptive behavior is a type of behavior that is used to adjust to another type of behavior or situation. This is often characterized by a kind of behavior that allows an individual to change an unconstructive or disruptive behavior to something more constructive. These behaviors are most often social or personal behaviors. For example a constant repetitive action could be re-focused on something that creates or builds something. In other words the behavior can be adapted to something else.” (http://en.wikipedia.org/wiki/Adaptive_behavior)
From this definition we can begin to think about the building as having a behavior which can change over time based on interaction with its environment. We can conceptualize the behavior of the building as constructive or unconstructive, and based on these definitions the building could adjust unconstructive behavior towards more productive behavior over time.
“Neural adaptation or sensory adaptation is a change over time in the responsiveness of the sensory system to a constant stimulus. It is usually experienced as a change in the stimulus. For example, if one rests one's hand on a table, one immediately feels the table's surface on one's skin. Within a few seconds, however, one ceases to feel the table's surface. The sensory neurons stimulated by the table's surface respond immediately, but then respond less and less until they may not respond at all; this is neural adaptation.” (http://en.wikipedia.org/wiki/Neural_adaptation)
This definition of adaptation begins to look at an understanding of neural networks and how they respond to various stimuli. Neural networks (NNs) and artificial neural networks (ANNs) will be discussed further below.
With a general understanding of the etymological root of the term adaptation and its historical use within the natural sciences as well as computer science, it is necessary to address certain topics taken directly from computer science in order for an architectural systems to embody what we would call “adaptivity.” These concepts include: intelligent agents, multi-agent systems, machine learning, etc...
“In artificial intelligence, an intelligent agent (IA) is an autonomous entity which observes and acts upon an environment (i.e. it is an agent) and directs its activity towards achieving goals (i.e. it is rational). Intelligent agents may also learn or use knowledge to achieve their goals. They may be very simple or very complex: a reflex machine such as a thermostat is an intelligent agent, as is a human being, as is a community of human beings working together towards a goal.” (http://en.wikipedia.org/wiki/Intelligent_agent)
The correlation of this definition to architecture is made obvious through the use of a thermostat as an example. Considering both the building systems and human occupants as agents links us directly back to the systems theory definition of adaptation from above, and we can consider the building to be a multi-agent system (see below for definition) composed of human and non-human agents.
Agent Oriented Programming
“Agent-oriented programming is a fairly new programming paradigm that supports a societal view of computation. In AOP, objects known as agents interact to achieve individual goals. Agents can exist in a structure as complex as a global internet or one as simple as a module of a common program. Agents can be autonomous entities, deciding their next step without the interference of a user, or they can be controllable, serving as a mediary between the user and another agent.” (http://en.wikibooks.org/wiki/Computer_Programming/Agent_Oriented_Programming
Agent oriented programming offers a code level approach to embedding intelligent behavior into non-human architectural agents.
“A multi-agent system (MAS) is a system composed of multiple interacting intelligent agents. Multi-agent systems can be used to solve problems which are difficult or impossible for an individual agent or monolithic system to solve. Examples of problems which are appropriate to multi-agent systems research include online trading, disaster response, and modeling social structures.” (http://en.wikipedia.org/wiki/Multi-agent_system))
Multi-agent systems have been extensively explored within the design stage of architecture and urbanism projects in order to simulate possible future situations which the architecture might encounter. (For an example of this type of multi-agent simulation at an urban scale see: http://vimeo.com/9158614 (‘Voxopolis’ city-engine by Dino Rossi, Jeanette Kuo & Dominik Zausinger, 2009) While there are many examples of MAS’s used in the design stage, instances of exploring the potential of MAS’s in built architecture are few and far between.
“Machine learning, a branch of artificial intelligence, is a scientific discipline that is concerned with the design and development of algorithms that allow computers to evolve behaviors based on empirical data, such as from sensor data or databases.” (http://en.wikipedia.org/wiki/Machine_learning)
Machine learning offers the structure through which to program adaptive behavior into buildings. While it is not romantic to think of buildings as machines (although many have and do) it can be a constructive way to embed adaptive behavior into buildings.
Artificial Neural Networks
“The term neural network was traditionally used to refer to a network or circuit of biological neurons. The modern usage of the term often refers to artificial neural networks, which are composed of artificial neurons or nodes. Thus the term has two distinct usages:
Biological neural networks are made up of real biological neurons that are connected or functionally related in the peripheral nervous system or the central nervous system. In the field of neuroscience, they are often identified as groups of neurons that perform a specific physiological function in laboratory analysis.
Artificial neural networks are composed of interconnecting artificial neurons (programming constructs that mimic the properties of biological neurons). Artificial neural networks may either be used to gain an understanding of biological neural networks, or for solving artificial intelligence problems without necessarily creating a model of a real biological system. The real, biological nervous system is highly complex and includes some features that may seem superfluous based on an understanding of artificial networks.” (http://en.wikipedia.org/wiki/Neural_network)
Clearly we are not dealing with biological neural networks when we are programing building systems, none the less it is important to understand the similarities and distinctions between these two types of neural networks in order to use the language of either appropriately. Artificial neural networks are a key component of many machine learning algorithms and are therefore necessary to understand in a bit more detail.
“A genetic algorithm (GA) is a search heuristic that mimics the process of natural evolution. This heuristic is routinely used to generate useful solutions to optimization and search problems. Genetic algorithms belong to the larger class of evolutionary algorithms (EA), which generate solutions to optimization problems using techniques inspired by natural evolution, such as inheritance, mutation, selection, and crossover.” (http://en.wikipedia.org/wiki/Genetic_algorithm)
Like ANNs, GAs play a key role in many machine learning algorithms and it is also necessary to understand them in some detail in order to discuss the various ways in which a building can adapt. Similar to Multi-agent systems, GA’s have been extensively explore in the design stage of architecture, but are relatively unexplored in relation to built architecture.
Back to Architecture
The “intentionality” of buildings.
As buildings embody increasing intelligence and gain the capacity to adapt to increasingly dynamic states, how do we describe the actions of buildings? At this point I would reference McCarthy’s description of the use of terminology such as “belief”, “desire”, and “intention” regarding machines:
“To ascribe certain beliefs, knowledge, free will, intentions, consciousness, abilities or wants to a machine or computer program is legitimate when such an ascription expresses the same information about the machine that it expresses about a person. It is useful when the ascription helps us understand the structure of the machine, its past or future behavior, or how to repair or improve it. It is perhaps never logically required even for humans, but expressing reasonably briefly what is actually known about the state of a machine in a particular situation may require ascribing mental qualities or qualities isomorphic to them. Theories of belief, knowledge and wanting can be constructed for machines in a simpler setting than for humans and later applied to humans. Ascription of mental qualities is most straightforward for machines of known structure such as thermostats and computer operating systems, but is most useful when applied to entities whose structure is very incompletely known.” (McCarthy 1979 http://www.agent.ai/doc/upload/200406/mcca79_1.pdf)
Per McCarthy’s description, I would argue for the use of terms such as belief, intention and desire in cases where the use of such terms facilitates the ease of discussion and understanding of increasingly complex building performance. This opens a method of discussing adaptive architecture not in what seem to many the opaque languages of computer science or biology, but in a language familiar to architects and the general public, assuming of course that use of such terms is grounded in general understanding of the origins of the language used.
The field of Adaptive Architecture is in its infancy and it is important to clarify which terms are appropriate to use when so that we may discuss in clear language the behavior and potential of architectures with the capacity to adapt. As part of the Assistant Chair for Architecture & Sustainable Building Technologies at the ETH Zürich, the research agenda being developed within the Adaptive Systems Lab will be focused on adaptive behavior in built architecture in relation to issues of sustainability. This does not mean that there are not other approaches to Adaptive Architecture which might have different agendas. There is a lot of interesting work being done in the field of adaptive architecture ranging from installation art to building control systems and we look to these works for knowledge and inspiration, in the same way that we look to other fields in the natural and social sciences.
In the end there are three basic requirements for an architecture to be truly adaptive in our opinion: The first requirement is that it must have the capacity to perceive its environment, the second requirement is that it must be able to act upon its environment, and the third requirement, this is what separates adaptive architecture from other categories such as kinetic and responsive, it must have the capacity to reflect on effect of its actions in order to adjust its actions in the future. If an architecture embodies these three capacities it can be said to be adaptive. What the capacities perceive, what actions they take and how they are reflected upon will necessarily change from project to project, but these three categories will be present in some form.
Note: This text is a work in progress, and will be further developed and changed as research continues within the ASL.
All Rights Reserved.