Critical Practice, a cluster of artists, researchers and academics hosted by Chelsea College of Art and Design has recently commissioned Ola Wasilkowska and Michal Piasecki to design a temporary spatial intervention for their Parade event. The event took place on 21-23 May, 2010 at Rootstein Hopkins Parade Ground in London.

The structure, called Assemblage in Public compromises computational design strategies with a low-tech assembly of repeatable components on site. The project is rooted in an idea that a repetition of a few simple rules is able to construct complex geometry, which can host different functions.

The core structure of the intervention was obtained through multi objective optimization with genetic algorithms. The optimization was conducted with a purpose built software written in Processing. The goal was to distribute the components according to a predefined “shadow map” and to keep them structurally stable at the same time. 4320 crates where used in total. Below are some of the individuals from one of the middle generations.

After the core was erected according to the plans extracted from the optimized form, everyone was invited to contribute to the intervention by adding on furniture. These where meant to be integrated and dissolved in the structure, so that discovering them may be surprising, but at the same time when they start to be used, they always prove comfortable. For this part of components assembly, there was no pre-defined layouts, so the structure started to crawl in diverse directions.

The low-tech crate component enabled the mixture of an evolutionary search for an optimal form with collective intelligence to create an attractor intervening the everyday use of the public.

Images: courtesy of Marsha Bradfield and Neil Cummings from Critical Practice.

“A Redefinition of the Paradox of Choice”, a paper co-authored by Michal Piasecki and Sean Hanna, was accepted for the Design Computing Cognition conference 2010 (DCC’10).

In the paper builds upon Barry Schwart’z definition of the paradox of choice. The abstract outlines our proposition of reconsideration of the paradox and positions the Breeding Objects Experiment 1 in this context:

Abstract: Barry Schwartz defined the paradox of choice as the fact that in western developed societies a large amount of choice is commonly associated with welfare and freedom but too much choice causes the feeling of less happiness, less satisfaction and can even lead to paralysis. The paradox of choice has been recognized as one of the major sources of mass confusion in context of the B2C online mass customization. We propose to redefine the paradox of choice with an emphasis on the meaning of choice in conjunction with the amount of available options, rather than just the quantity of choice. We propose that it is the lack of meaningful choice, rather than an overwhelming amount of choice, that can cause customers’ feelings of decreased happiness, decreased satisfaction and paralysis. We further propose that since users themselves are often not able to explicitly define what constitutes a meaningful choice, the task they face belongs to the category of ill-defined problems. The challenge for mass customization practitioners is thus not to limit the scope of choice, as has been suggested in previous literature, but to provide users with choice that is relevant to them.

We further discuss two computational approaches to solving problems related to the redefined paradox of choice in the context of the B2C mass customization. The first is based on recommender systems and the second is an implementation of artificial selection in genetic algorithms. We present findings of an empirical comparison of genetic algorithm and parametric product configurators. We find that the genetic algorithm tools, which allow users to move through a solution space by recognition of meaningful options rather than their definition, appear to be more popular among the users when it comes to browsing through solution spaces with larger number of dimensions.

This was confirmed by more than 300 sessions with the configurators, which took place during last three weeks. Parametric configurators with 4 and 18 customizable parameters (4P and 18P) where more popular than their genetic algorithms counterparts (4GA and 18GA), but genetic algorithm was a much more popular mean of driving the 37th parameters definition (37GA turned out to be more popular than 37P). I’m currently writing these findings up into a paper and some further, more in-depth analysis of the results will follow.

For the next iteration of the experiment, Federico Weber has generously offered to share his customizable table design with me. The work called Xylem is based on Voronoi diagram and will become a phenotype in Breeding Objects Experiment 2. Federico has already developed a parametric configurator for Xylem. It’s downloadable from here.

Breeding Objects Experiment 01 is my first experiment with configurators for mass customization of products over the Internet, conducted as part of PhD research at the Bartlett School of Graduate Studies. The experiment is a comparison of two types of configurators: “parametric” configurators and genetic algorithms configurators. In parametric configurators users have to specify the values of each of the parameters explicitly, while in genetic algorithms configurators they are not required to do so. Thanks to artificial selection, they can instead navigate through solution space by breeding objects, that is manually selecting parents of next generations.

The experiment is only about the patterns of interaction with the tools available, so you will not be asked to fill in any questionnaires. Once you start interaction with the tools, you will be able to navigate between them using links on top of each page.

Evolve rings where commissioned and are distributed by Fluid Forms. I have co-designed them together with Krystian Kwiecinski and they are available for purchase since today. They where designed with an aid of an artificial selection in genetic algorithm (GA). I have developed the phenotype together with Krystian and then wrote a purpose-built GA tool for breeding, rather than designing products in Processing. We will convert the GA tool into a configurator for mass customization of the rings. It should become available on Fluid Forms site early in 2010. Below is an explanation of the roots of evolutionary approach in design, which I wrote for the purpose of the launch.

Roots of the Evolve ring generative design process

Today Darwinian model of evolution by natural selection is widely acknowledged. It assumes that a new generation of individuals within a given specie inherits properties of these individuals from the previous generation, which where the most fit to survive. Fitness to survival is always evaluated within a given environment. It is not objective, but a subject to particular conditions which the individual happened to find itself in.

On a metaphorical level, theorists has refereed to evolution of culture, societies and products that they develop. Philip Steadman’s “The Evolution of Designs” from 1979 is a great roadmap of evolutionary metaphor in reference to any kind of artifacts, including designer objects and architecture. Bashford Dean’s study of historical evolution of helmet design or Lane-Fox’s and Pitt-Rivers’ tracing of evolutionary relationships of Australian weapons are just some of the examples discussed in Steadman’s influential book.

With the dawn of computing, evolution could become an applied tool used for seeking answers to particular problems rather than a conceptual model only. Survival of the fittest has been a source of inspiration for developers of “evolutionary algorithms”, which enable navigation through vast spaces of potential solutions in any imaginable field from marketing to transportation. They also help find solutions which are a best fit for given criteria, even if these are very complex and somewhat contradictory. The art world was quick to adapt the notion of evolution and survival of the fittest, but engineering was as well. In the art world William Latham imagined artists growing artworks rather than creating them already in 1985.

Later his renowned collaboration with IBM’s Stephen Todd led to development of the software called “Mutator”. The software literally enabled Latham to breed his sculptures just like the gardener breeds plants. In this case the individual most fit to survive is the one the artist likes the most. The possibility of breeding jewelry is also mentioned in Latham’s and Todd’s contribution to Evolutionary Design by Computers edited by Peter Bentley.

This is in contrast to application of evolution in engineering, where aesthetics do not have a significant impact on the validity of the solution. What is important in this case is performance. An example is the work of dr Peter Bentley and Stephen Manos on optimization of Holey optical fibres. These fibres are able to transmit light on large distances and their capacities are greatly enhanced by small air holes running through their entire length. Question was what shape, size and number of the wholes provides a fibre with best transmission capacity? Looking for an answer with evolutionary computation likely took millions of solutions to develop, but as the process was automated the time lapse was not so big and the results where rewarding.

Evolve ring brings evolution to the design world, not as a metaphor, but as a design tool. In design, aesthetics plays a crucial role, but the design has to be valid from a manufacturing stand point as well. This is why an algorithm used by Michal Piasecki and Krystian Kwiecinski is an amalgam of the GA from the art world and from the engineering world. Michal and Krystian developed a parametric phenotype first. Later Michal unleashed the process of evolution of jewelry and guided it towards the designs, which became the Evolve rings designer’s series from Fluid Forms. These pieces are breed, not designed.

John Frazer’s “An Evolutionary Architecture” from 1995 is an obligatory read for anyone interested in the amalgam of evolutionary computation and design. In other words anyone interested in the potential of “breeding designs” using evolutionary algorithms (genetic algorithms belong to this category). I was very happy to see it available online for free from the AA website.

I recently wrote an article about Sean Hanna’s and Siavash H. Mahdavi’s research into microstructure optimization for Architektura – Murator – a polish architectural design magazine.

Sean Hanna is a RCUK Academic Fellow in Space and Adaptive Architectures based at the Bartlett School of Graduate Studies of UCL. His microstructure optimization is essentially a design of an optimized material to be produced with SLS. It is conceptually grounded by an observation of the behavior of the cells of wood, which represent a degree of local intelligence. They adapt (optimize) their form according to their position in the overall structure. Microstructure is also composed out of cells, which are optimized in two ways (or for two kinds of conditions).

optimization for manufacturing constraints

First kind of optimization is an multiple-objective optimization for the digital manufacturing technique constraints. A genetic algorithm is used in order to optimize the shape of the elements composing each cell. The goal is to be efficient in terms of the amount of material used but still be able to produce the object with an SLS machine. Underneath you can see an example of a population of solutions.

structural optimization

Second type of optimization is structural optimization. Forces in a specific shape which is about to be produced in microstructure are analyzed. Then the thicknesses of the cells are adjusted accordingly allowing for even greater efficiency of the material. Support vector machines are used in this process. In opposition to the optimization for manufacturing constraints, this type of optimization has to take place every time new object is about to be produced out of microstructure.

sample – a cantilever

A cantilever was chosen as a sample of microstructure. The deflection of the microstructure cantilever is several times smaller than the one of a standard engineered structure (image above). The sample has only 2 cm but I believe the potential of the technique is enormous.

images: courtesy of Sean Hanna

I have focused on the engraving feature of 500 speedy laser cutter. The software driving it has a capacity to read raster images and convert them into engravings. Effectively it has 16 depths of engraving, so I decided to write a cellular automata with 16 possible states of each cell. States vary between one another by a fill() value in Processing. As you would assume, the fill value goes from black to white (black is the deepest engraving).

Here are some first tryouts. You can see first row, generated randomly and the rest of the image following certain rules. Every cell in the next row is “looking at” three cells above itself. As in every CA the inheritance from previous cells takes place based on CA rules.

In the image above I have assigned the CA rules arbitrary. Next step was to drive a CA with a GA (genetic algorithm). I have used a similar interface as in the case of Breeding Objects. The difference being that only artificial selection was available to the user. It means that the only way for the user to create next generation was to manually select individuals to breed from the current generation. This choice is made purely based on users’ aesthetic preferences.

In the first iteration of CA driven by a GA only the CA rules where inherited by next generation but the states of the first row where random in every new individual. This approach proved not to be very successful, since a very small change in the first row’s states could cause a considerably big change in the overall image. In the final software both CA rules and the states of the first row evolve and are inherited by next generations.

You can see some of the CA images which got engraved in Corian during the session (images of the physical objects coming soon).

UPDATE: Few pictures of the laser cut.

The video above is the final presentation of my second term programming project at the Msc Adaptive Architecture and Computation at the Bartlett, UCL. Underneath you can find a detailed description of the software as well as some photos of phenotypes fabricated for the 2nd term final exhibition which took place at Bartlett’s Wates House on Monday.

You can also take a look at previous posts about this project here and here, although I’m afraid that the explanations there will overlap with the one underneath.

external embryogeny

External embryogeny is a type of GA in which fitness function is pre-defined. In external embryogeny only the populations of solutions are subjects to evolution. It’s opposite is an implicit embryology GA – in which fitness objective evolves together with populations of solutions.

In the following month I’m going to write a paper about “breeding objects” software.
The research question which I hope to tackle in it is whether it is possible to harness an explicit embryology GA as a tool providing design solutions for both well-defined and ill-defined design problems. Natural and artificial selection is incorporated into the software in order to enable such possibility.

ill-defined and well-defined design problems

Well-defined design problem is a quantifiable problem. Budget constraint can be an example of such problem. Performance of the objects – for example the the amount of exposure to the light which they receive might be another example of it.

Ill-defined design problem in a problem which is not quantifiable. Aesthetics is a classic example. There has been many interesting attempts to research general patterns in aesthetic choices during design process, but, to date, every attempt to extract a general rule from the findings failed to some extend. The incorporation of natural and artificial selection into a GA might makes generalization obsolete.

phenotype generative procedure

The advantage of GA is that it can accommodate quite complex generative procedures. The final generative procedure is the 4th one which I was experimenting with. This proves that different procedures can be plugged to the same GA as long as their fitness objective is similar (this is explained in natural selection paragraph). The parameters of final procedure are the size of the object, resolution of the point grid as well as grid’s distortions caused by set of detractors and attractors. Finally range and strength of every attractor and detractor is also a parameter.

natural selection

Natural selection provides an answer to well-defined problems. In other words it is an optimization tool. In case of “breeding objects” software the problem (the fitness objective) is the cost of the object. Since the fabrication method is 2d laser cutting – the cost of the object consists of length of all of the cuts (which determines laser cutting time) as well as the area of the material.

During the natural selection step the software itself determines which individuals it should breed in order to create next generation. The choice is made based on the fitnesses of all of the individuals (the difference between their cost and the fitness objective).

The population quite quickly converges to a fit design solution. However if the process of natural selection continues the it is able to converge into another design solution. This solution is equally fit as the previous one (it’s cost is roughly the same) but it is very different visually.

artificial selection

The convergence from one fit design solution into another during natural selection proves that aesthetics is an ill-defined problem. This is where artificial selection comes in.

During the artificial selection step the user makes the choice about the individuals which become parents of next generation. User can choose just one individual or the whole population as the ones to breed.

The user is informed about the fitnesses of all of the individuals. Their choice can but does not have to be determined by this information. Additional information during this step is provided by 3 viewports enabled in the software: perspective, elevation and shading ratio (the amount of light which the surface underneath every individual receives).

natural & artificial

If the user will deliberately choose individuals with low fitness he or she is likely to loose the achievements of natural selection. This means that the next population can consist of less fit individuals than the previous one. It is not clear however, that it is always the case. What I’m actually going to investigate in the paper is whether that depends on some pre-determined parameters of both natural and artificial selection steps. Those parameters might for example be the percentage of mutation of each individual’s genotype or a range of fit individuals from which parents of next generation are chosen during natural selection.

Now, I have incorporated an artificial selection step into it. I took a look at what William Todd and Stephen Latham propose in “Evolutionary Design by Computers” (edited by Peter Bentley). They have suggested that, while simulating the process of evolution, there is no reason to limit the number of parents of future generations to only two. The software I’m writing already allows for a multiple choice of parents.