Agency, Art, and Architecture

Permalink

When we examine the field of swarm intelligence, we must ask ourselves if what we are studying has any real-world applications. The topic of nest building relates to architecture in many ways. When building nests, swarms have: a system for creating structure, designation of duties, experience with materials, and system of delineating social hierarchies. If insects can coordinate their behaviors in order to build these highly complex architectures, how can we, who have a much higher capacity for analysis and problem solving, emulate this behavior and even take it to a much higher level? The goal is to understand how the insects stimuli are organized in space and time and use this understanding to apply a system to organizing our built environment. Today, most humans live in an extremely complex and stimulating environment. We can use this to our advantage. Studies have found that the most populous termite communities generally have the most complex nests. This shows that the complexity of our environments can be an advantage. The success of these large termite communities is due to their ability to self organize. “A social insect colony is a decentralized system composed of cooperative, autonomous units that are distributed in the environment, exhibit simple probabilistic stimulus-response behavior, and have access to local information.” One of the main ingredients of SO is multiple interactions, these may be direct or indirect. For wasps, “building decisions are made locally, influenced by the existing configuration”, this is something that we can carry over into architecture. Responding to existing conditions and configurations, could lead to possibilities in form development and sustainable building practices.
Insects in these complex systems respond to stimuli that originates from their neighbors or their environment and respond with a simple positive or negative behavior response. This idea of agents or insects responding both directly and indirectly with their environments is known as stimergy. When observing the nest building of wasps, one is able to see stimergy in full effect. As the nests get bigger the options for where the next “pod” can be placed greatly increase, showing how the continually changing environment effects the insects decision making. These newly emerging options guide the wasps decisions as to where to continue to build. Where the wasps choose to add the new pod is not completely random. For example, they are more likely to add a pod where 3 adjacent walls are present rather than initiate a new row. These decisions lead to emergence of new forms.
This behavior can be modeled in genetic algorithms by informing agents to respond to a series of if-then decision loops. The first act consists of only a few exceptions. GA’s designed to emulate nest building often include agents that “move in a three-dimensional grid and drop elementary building blocks depending on the configuration of blocks on their neighborhood.” “The stimuli that initially triggers building behavior may be very simple but as construction progresses, stimuli is multiplied and becomes more complex along with the insects (or agents) responses to the stimuli”
Theraulaz & Bonabeau’s set up these rules for their simulation:
deposit of a brick by an agent depends solely on the local configuration of bricks
once bricks are deposited they cannot be removed’
all simulations start with a single brick
configure a microrule table
Some notes on the behaviors of these genetic algorithms:
“A genetic algorithm based on the fitness function is used to explore the space of architectures”
Non-coordinated algorithms tend to overlap and the architecture seems to grow in space without any coherence
Coordinated algorithms obtain more variation if more options are available at specified steps
artificial evolution is particularly important because it picks the most functional or aesthetic from the many forms produced by the GA’s
GA’s often produce structured patterns
Up next … patterns
If you want to do some more reading on the possibilities that mimicking insects provides, check this out…. http://www.archdaily.com/34235/the-termite-pavilion/
Sources:
Swarm Intelligence: From Natural to Artificial Systems by Eric Bonabeau, Marco Dorigo and Guy Theraulaz. (1999)
“The Termite Pavilion | ArchDaily”, n.d., http://www.archdaily.com/34235/the-termite-pavilion/.

Permalink

While researching swarm intelligence I stumbled across an instillation artist named Federico Diaz who I thought was relevant and particularly interesting. “Since the early 1990s, Díaz has focused on art which combines software generated shapes and sound with work in space. Multimedia installations and realizations related to biomorphic architecture and site-specific projects are central to his domain, and he commonly engages audiences through active participation in his art.”
“In his instillation Geometric Death Frequency—141, he transforms a photograph into pure data, and modulates it using analytical and fluid dynamic modeling techniques, finally rendering the data stream into a three-dimensional sculpture using state-of-the-art computer-aided manufacturing methodologies. The new work thus combines elements of photographic manipulation, data analysis, and computer programming, utilizing new techniques to produce a sculpture completely untouched by human hands. In this instance Federico has ‘deconstructed’ a photographic image of our entrance courtyard. Taking the digital bits and bytes which define the location, depth, and color of the pixels from the photographic image, Federico rearranges them thru the artful application of the laws of fluid dynamics. The three-dimensional black spheres that make up this work are in some ways distant relatives of the pixels of that original, two-dimensional photograph, but instead of being transcribed into a flat image as micro dots of ink on paper, here they well up into great waves that crash against the very walls of their original subject matter.”

Permalink

Architecture today is dominated by control, and programming. Recently, there has been a call for “intelligent systems” that respond to their environment and are developed from principles of emergence and distributed functioning. In Neil Leach’ s article Swarm Tectonics: a manifesto for an emergent architecture, he discusses the bifurcation of the architectural field throughout history. He separates the two associating one primarily with aesthetics and the other with structure. The former “tends to ‘impose’ form on building materials, according to some pre-ordained ‘template’. (And here one immediately thinks of the role of ‘proportions’ and other systems of visual ordering.)” and the later concentrates on structural and programmatic requirements that allow forms to emerge based on the organization of the buildings interior.
Leach goes on to discuss how advances in technology occur over time. New forms emerge and are perfected through a long process of trial and error. Now, it is the responsibility of the architect to harness this new technology (computation through computerization) and attempt to speed up the emergent process from hundreds of years to a matter of minutes through computer simulations and rapid prototyping. This process allows designers to “engage with the principles of structural engineering … not as some practical afterthought but as a vital component folded into the whole conceptual process of designing.”
“One such example is the eifForm program, devised by Kristina Shea, (as shown above) that generates forms in a stochastic, non-monotonic method using a process of structural shape annealing. The ‘designer’ merely establishes certain defining coordinates, and then unleashes the program which eventually ‘crystallizes’ and resolves itself into a certain configuration. Each configuration is a structural form which will support itself against gravity and other prescribed loadings, and yet each configuration thrown up by the program is different. Such is the logic of a bottom-up, stochastic method.”The potential of this technology does not stop with generation of new structural forms but has the potential to recognize optimal solutions for construction with a limited set of resources. This can become an important design tool in less affluent parts of the world where resources and technology are scarce.

Permalink

Swarm Intelligence is a sub-filed of Genetic Programming. Swarm intelligence refers to the emergent collective intelligence of a group of agents. It is often seen in nature through schools of fish, flocks of birds, swarms of bees, and many other complex animal and cellular systems. Researchers are looking to these systems because of their perceived complexity and their ability to be broken down to a series of simple actions that make up a more complex system. This phenomenon is known as recursion. In these systems, autonomy, emergence, and distributed functioning rule.
This can be seen repeatedly in insect colonies. Although insects possess a certain amount of intelligence, they are not intelligent enough to explain the complexities that make up the larger social colony. To understand how these colonies work, one must consider how to “connect individual behavior with collective performance.?” Many of the activities of these insects are self organized, based on a set of positive and negative feedback, but the colonies are complex enough to continue to function even though some individuals may fail to perform their tasks. Because of the insects limited cognitive abilities, it becomes feasible to model agents that mimic their behavior. Modeling these multi-agent systems can “help design artificial distributed problem-solving devices that self organize to solve problems.
*Source:
“Eric Bonabeau, Marco Dorigo and Guy Theraulaz: Swarm Intelligence - From Natural to Artificial Systems”, n.d., http://jasss.soc.surrey.ac.uk/4/1/reviews/kluegl.html.

Permalink

A genetic algorithm is a “system” that mimics the process of evolution and natural selection. They are often used to generate “optimized” solutions using techniques that occur in nature such as selection and mutation. There are four things that are always present in GA simulations:
Population - represents possible solutions to the problem
Fitness Function - a procedure to discriminate good from bad solutions to a problem
Selection Function - uses values set by the Fitness Function to select which members of the Population will be incorporated into the next “generation”
Mutation Operator - produces members of the next “generation” by copying existing members and producing “variation” in the Population
GA’s are being applied to architectural design for two reasons:
Potential for Design Optimization
As a form generating tool based on the concept of Emergence
Design Optimization is used to enhance building performance by testing many variables in search of the most efficient solution to the problem. It may be used in structural planning to minimize weight and material cost (as seen in the Water Cube), in lighting, energy, programming (modeling optimum adjacency), and environmental performance (which may be modeled using AutoLISP which allows results to be visualized as AutoCAD drawings). Design Optimization problems contain 3 things:
Objective Function - what needs to be minimized or maximized
Designation of Variables - affect the values of the objective function
Determine Constraints - allow the design variables to have a set of values
Using genetic algorithms as a form generating tool is not an exact science. GA’s generate forms based on a set of rules that emulate natural systems. It becomes the responsibility of the architect to assign an appropriate set of rules to generate the form of the building. The architect must keep in mind that the building is not a living system but consists of many interconnected systems that must be considered when defining the parameters of the genetic algorithm. In genetic programming Recursion - functions that result from the composition of simpler ones, is a common occurrence. This phenomenon is commonly depicted as a tree. This tree consists of Elementary Functions that form the branching points of the tree, Variables and Constraints that represent the leaves of the tree and are used by Elementary Functions as inputs, and Fitness which measures the degree of distance to the target. An architect may program many simple functions that describe the many systems of the building and through recursion on overall design scheme may be established.
“Architectural design might be aligned with neither formalism nor rationalism but with intelligent form and traceable creativity.”3
*Sources:
Manuel De Landa, Philosophy and Simulation: The Emergence of Synthetic Reason (Continuum International Publishing Group, 2011).
http://www.generativeart.com/on/cic/papersGA2007/09.pdf“
Expressive form: a conceptual … - Kostas Terzidis - Google Books”, n.d.,

Permalink

What is an algorithm?

An “algorithm is a computational procedure for addressing a problem in a finite number of steps… it is the systematic extraction of logical principles and the development of a generic solution plan.” Algorithms can be powerful tools because of there ability to work much like the human mind. They are able to provide new solutions to problems that t humans may not be able to calculate or visualize on their own. “Inductive algorithms can be regarded as extensions of human thinking because of their ability to explore generative processes or simulate complex phenomena.”

Computation is the method algorithms to produce these complex systems.Computation is the procedure of calculating while computerization is the act of entering, processing, or storing information in a computer.Computation is a method of exploring ill-defined processes. It involves problem solving based on a set of rules that mimics the human intellect but is able to perform functions more efficiently. Computerization deals less with the abstract and more with well defined problems. It involves digitization which is the “conversion of analog information into digital information.”

As discussed in a previous post walking blindly, “Algorithms are encapsulations of processes or systems of processes that allow one to leap and venture into the world of the unknown, whether natural or artificial. They are not the end product, but rather a vehicle for exploration.”

*sources:

“Expressive form: a conceptual … - Kostas Terzidis - Google Books”, n.d.

Permalink

Now the big question… How do I use agency to generate a system and later a thesis? I have always been interested in how natural systems influence architecture so I thought I would start there. Exploration of a natural system to generate structure and form. In reading a 2006 issue of AD magazine, Collective Intelligence in Design, I came across an article, Collective Cognition: Neural Fabrics and Social Software. This article discusses “how does an organism produce meaningful information about the environment from an uninterrupted stream of data?” This article caught my eye because I had previously contemplated researching how people experience space without vision. Incorporating brain mapping as my “natural system” seemed to be a way to incorporate this earlier idea and agent-based modeling.
“An analysis of the neurobiological system provides a procedural model, or abstract machine, from which certain organizational principles may be captured… to study these complex connections, scientists began to study the brains connections through dynamic process based models such as network typologies.”
Some characteristics of the brain’s neural fabrics are as follows:
function of the system depends not only on the elements but also on the way they are connected
elements alter form in response to activity
self organizes through associations within limited hierarchies
connections are point-to-point but continually changing
most interactions are local but there is additional connections globally
transmissions are non-linear and recursively loop back to previous neurons as well as themselves

“Neurological organization can be summarized as a distributed meshwork characterized by emergence of synthetic reason.”After discussing this proposal with my tutor, is was suggested I start with a simpler subject or natural system such as swarm intelligence. Although I was disappointed at first, I think he was right. The brain is an extremely complex system that would be quite a challenge to code.

Permalink

What is agent-based modeling?

“In agent based modeling (ABM), a system is modeled as a collection of autonomous decision making entities called agents. Each agent individually assesses it’s situation and makes decisions on the basis of a set of rules. Agents may execute various behaviors appropriate for the system they represent.” -Bonabeau Agent-based modeling: Methods and techniques for simulating human systems

They key to agent based modeling is realizing how to use the system and realizing it’s benefits and limitations. ABM is now being used in many different fields such as medicine, finance, retail, and marketing but what makes it particularly interesting to Architects are: it’s “ability to capture emergent phenomena” which can be a useful design tool, (“emergent phenomena result(s) from the interactions of individual entities. By definition, they cannot be reduced to the system’s parts: the whole is more than the sum of its parts because of the interactions between the parts. An emergent phenomenon can have properties that are decoupled from the properties of the part” -Bonabeau); it “provides a natural description of a system”, something that architects are always striving to do- rationalize natural systems and apply them to the built environment; and finally, it “is flexible” able to adapt to many variables and be quickly manipulated to allow architects to immediately visualize how changes in variables can effect the overall system.

Permalink

Walking blindly

Throughout my time at architecture school I have been criticized for one thing more than anything else… creating form for form’s sake. Learning Processing will enable me to experiment with form through a generative process rather than just creating it abstractly. However, I am still concerned about how the rest of the faculty will receive my project. Reading “The Autopoiesis of Architecture” by Patrik Schumacher reminded me that school is the time to experiment and learn, not to generate “commercial designs”. Referring to educational institutions position in the Architectural avant-garde movenent, Schumacher states that schools should be an “incubation chamber that shield(s) radical experimentation from the pressures of mainstream feasibility.”

My whole life I hay been asked “why.” “Why are you doing that,” and “where is this going?” For the first time, I don’t have to have an answer. Strict rules have long governed the architecture profession. During the Renaissance it was all about form and proportion with functionality taking a backseat and Modernism preached program as a generator of form. Now, generating form, or program, or both through agent based modeling will allow me to slowly discover my goals rather than work from a top down approach. ”The freedom to post-rationalize is greatest where no specific problem is posed from the form - the only requirement being that a form -function relationship is established at the end. This is only possible within academia.” It is a bit unnerving not knowing where this process will take me. I feel like I am walking blindfolded off a cliff and hoping there is a net below that will catch me. I have been reassured by my professor and once again, Schumacher that “The potential of an architectural experiment is often discovered in retrospect. Goals are the outcome rather than the starting point of research.”

Permalink

The big decision

I arrived at school this year feeling behind schedule before the first day of class. Not the best way to start. When we all assembled to pick our research professors I was nervous because I had nothing. What was I going to pull out of my ass before the first class. I tried to think how I could turn graffiti into a thesis topic to no avail so I sat through the professor’s presentations looking for a match or someone who could send me in the right direction. Gabriel’s presentation spoke to me. I have been interested in parametrics and scripting for quite a while but I didn’t think there was any staff at this school who were knowledgeable enough to help me with these subjects and I thought the topic was too daunting to try to tackle on my own. Gabriel and I had our first meeting and I learned that I would be studying algorithms. I was a little scared but excited for the possibilities.

Permalink