Conference PaperPDF Available

Evolutionary Design of Housing: A template for development and evaluation procedures


Abstract and Figures

Evolutionary design is an approach that evolves populations of design variants through the iterative application of a set of computational procedures. This paper proposes a template and set of techniques for creating the development and evaluation procedures. The template defines a clear structure for the procedures, while the techniques provide specific strategies for generating models and handling constraints. A demonstration is presented where the template is used to create development and evaluation procedures for a large complex residential housing project.
Content may be subject to copyright.
M. A. Schnabel (ed.), Cutting Edge: 47th International Conference of the Architectural Science Associa-
tion, pp. 197206. © 2013, The Architectural Science Association (ANZAScA), Australia
A template for development and evaluation procedures
National University of Singapore, Singapore,
Abstract. Evolutionary design is an approach that evolves populations
of design variants through the iterative application of a set of compu-
tational procedures. This paper proposes a template and set of tech-
niques for creating the development and evaluation procedures. The
template defines a clear structure for the procedures, while the tech-
niques provide specific strategies for generating models and handling
constraints. A demonstration is presented where the template is used
to create development and evaluation procedures for a large complex
residential housing project.
Keywords. Evolutionary design; generative modelling; constraint
handling; decision chain encoding; point block housing.
1. Introduction
Evolutionary design (Frazer 1995, Bentley 1999, Caldas 2001, Bentley and
Corne 2002, Janssen 2004) is an approach that evolves populations of design
variants through the iterative application of a set of computational proce-
dures. The development procedure generates design variants, one or more
evaluation procedures assess the performance of design variants, and the
feedback procedure drives the evolutionary process by applying selective
pressure to the population. The feedback procedure applies selective pres-
sure by ensuring that design variants with low performance scores are more
likely to be killed, while design variants with high performance scores are
more likely to survive, and to be selected for reproduction.
This paper will focus mainly on the development and evaluation proce-
dures. Section 2 describes a template for creating such development and
evaluation procedures, section 3 presents the demonstration of the applica-
tion of the template, and section 4 briefly draws conclusions and indicates
avenues of further research.
2. The Development and Evaluation Procedures
The development procedure generates a phenotype, which is a design vari-
ant. One or more evaluation procedures generate a set of evaluation scores,
which are measures of performance for a design variant. Janssen and
Kaushik (2013a) proposed a template for development procedures. This pa-
per builds on this previous template, by expanding the scope to include both
development procedures and evaluation procedures. Figure 1 shows the pro-
cedures and sub-procedures of the proposed template.
Figure 1: The template for development and evaluation procedures
The development procedure starts with a model of the environment (e.g. the
site and surroundings) and generates a skeleton model of the design variant,
under the influence of a set of genes. The aim is to create development pro-
cedures that result in phenomes with sufficient design variability. Phenomes
for architectural and urban designs are typically both highly variable and
highly constrained. They are highly variable in the sense that there is no
fixed organisational plan, but instead entities can be organised in space in a
wide variety of ways. At the same time, these organisations are highly con-
strained by various rules delineating the validity of possible designs. This
type of phenome that is both highly variable and at the same time highly
constrained is described as having bounded variability.
In order to achieve bounded variability, a key challenge is effectively
handling constraints. In particular, two type of constraints need to be han-
dled: combinatorial constraints and geometric constraints. For combinatorial
constraints, a technique called decision chain encoding can be used (Janssen,
2004; Janssen and Kaushik, 2013b), while for geometric constraints, dynam-
ic solvers can be used. Together, these two sets of techniques form a simple
yet powerful toolkit for handling a wide variety of constraints.
The decision chain encoding technique structures the skeleton generation
process as a sequential chain of decision points. Each decision point involves
choosing one option from a list of options. The list of options is created by a
set of rules that generate options and then and then filter out options that vio-
late constraints. Note that for each decision, the total number of valid options
may not be known and may depend on the previous decisions.
The decision chain encoding technique can therefore be used to generate
configurations that adhere to a range of combinatorial constraints. However,
the resulting configurations may still violate other geometric constraints. For
resolving these geometric constraint violations, dynamics solvers can be
used. These dynamics solvers will, over a series of time steps, try to modify
the configuration in order to resolve any constraint violations. Depending on
the types of constraints, a variety of dynamic solvers can be used, such as
particle solvers, rigid body solvers, inverse kinematic solvers, and cloth
The evaluation procedure starts with a skeleton model and generates an
evaluation score for the design variant. As a side effect, the evaluation pro-
cedures also generate domain models required for analysis or simulation.
Since the skeleton model is a sparse model, data compensation techniques
need to be used in order to add the missing data.
3. Demonstration
In this section, the implementation of a development procedure and a set of
evaluation procedures for an example design schema are described. The
schema is for a residential housing development consisting of a set of point
In the design scenario, it is envisaged that a developer plans to build a set of
residential buildings with flats arranged around central cores containing cir-
culation and services, a typical typology referred to as a 'point-block'. Typi-
cal layouts of the individual flats are defined in advance but the positions
and heights of the point blocks and the number of flat types for each point
block can be varied.
The site is located in Singapore, with an area of 8.4 hectares and a plot
ratio of 2.0. In total, 1400 flats are required. Figure 2 shows the 7 flat types
(together with the required quota for each flat type), and the 4 block types.
Flats are always arranged around the core in pairs, sharing a common wall
and forming vertical stacks of flats of variable height (each between 6 to 12
floors high). At the ground level, each block type can accommodate a differ-
ent number of flats around the core (4, 6, and 8 flats). However, due to the
variable stack heights, the number of flats and the core may reduce as it goes
Figure 2: Flat types and block types
Figure 3 shows an example of a single block, together with a conceptual
section showing a level of car parking at the bottom covered by a landscaped
level on top. The blocks can be freely positioned on the site, with all blocks
being accessible either by car via the lower car park and by foot via the up-
per landscaped level. The upper level has greenery as well as swimming
pools and playgrounds. The aim is to optimise the configuration of point
blocks and flats so as maximise saleable value and at the same time maxim-
ise a number of window performance criteria (described in more detail be-
Both the development and evaluation procedures were created in the pro-
cedural modelling software SideFX Houdini. This software allows these
procedures to be defined visually using Visual Dataflow Modelling (VDM)
(Janssen and Chen 2011). In addition, the software also includes a wide
range of procedural modelling tools and dynamic solvers.
Figure 3: Conceptual section across the site
The development procedure uses a decision chain encoding technique in or-
der to handle the combinatorial constraints on the flat types and a dynamics
particle solver technique in order to handle the geometric constraints related
to point block positioning.
The point block configurations are generated using a combination of sim-
ple parametric modelling techniques and decision chain encoding tech-
niques. The genotype is structured so that there are a repeating set of 16
genes defined for each point block.
The process of positioning and orientating the block on site is performed
using some simple parametric rules. For positioning, the site area is mod-
elled as a UV surface, and two genes are used to define a UV position on
that surface. For orientation, the third gene is used to rotate the block be-
tween 0 and 360 degrees. This process of positioning and orientation ignores
possible collisions between blocks, as those will be resolved later using the
dynamics particle solver. The remaining 13 genes are used to form a single
block of flats through the decision chain encoding process. This process
takes into account various combinatorial constraints in choosing the appro-
priate flat types and stack heights so as to not overshoot the quota.
Since the blocks can be of variable height, the total number of blocks re-
quired to achieve the desired number of flats may also vary, with the maxi-
mum number of blocks set at 32 blocks. In order to handle this variability,
genotypes with redundant genes are used. Since there are a maximum of 32
blocks with 16 genes per block, the total number of genes for generating the
blocks is 480. However, due to the redundancy, some of these genes may not
be used. For example, if a block is of type 1 (4 flats per floor), then it will
only require 2 stack genes and 4 flat type genes, meaning that the other stack
and flat type genes are not used. Similarly, if the required number of flats
has been achieved with 30 blocks, then the last 32 (16x2) genes will not be
In addition the blocks, swimming pools and playgrounds are also added
to the site layout. These are positioned in the same way as the blocks, using
UV positioning genes. These additional programmatic functions are not di-
rectly evaluated, but they play an important role since they create open spac-
es between the blocks.
Figure 4: Dynamics particle solver repositioning particles on the site.
Once all the blocks, swimming pools, and playgrounds have been gener-
ated, they may be intersecting and overlapping. In order to resolve these is-
sues, a dynamics particle solver is used. For this solver, each block, swim-
ming pool, or playground is represented as a circular particle. For the blocks,
the radius of the particle is adjusted to fit the size of the block. The site
boundary is defined as a particle boundary and the particles are then posi-
tively charged so that they repel one another. These particles are then ani-
mated for 1000 frames, allowing the particle to reposition themselves, there-
by automatically resolving the overlaps between the blocks. Figure 4 shows
a set of frames from the animation.
The first step in the evaluation process is the generation of the domain spe-
cific models. The skeleton model resulting from the development procedure
is a sparse 2D skeletal model. For each block, the model contains a set of
polygons that represent the plans of the flat types tagged with attributes de-
fining the number of floors. A number of different domain specific models
are then generated from this 2D skeleton as inputs for analysis and simula-
tion. In addition to the domain models, visualisation models can also be gen-
erated in order to help designers evaluate other aspects of the design. Figure
5 shows the 2D skeleton model on the left, along with three other models
generated from this skeleton.
Figure 5: Different types of models generated from the 2D skeleton.
For evaluating saleable value, a domain model is generated that includes
only the floor plates of the individual flats. The calculation uses a simple
formula that adjusts the price per square meter according to the floor level.
For evaluating window performance, a domain model is generated that
only includes the outer building surface together with the living room and
bedroom windows of the flats. The window performance takes into account
certain site conditions, including a canal along one side of the site that is
treated as a desirable view, and a number of busy roads that are treated as
sources of noise pollution. For each window, three different criteria are con-
Maximisation of unobstructed view in front of the window, where 100% in-
dicates a completely unobstructed view of 50 meter radius.
Maximisation of views of the canal, where 100% indicates that the whole
stretch of the canal in front of the site is visible.
Minimisation of exposure to road noise, where 100% indicates that there is a
road directly in front of the window.
These three criteria could be treated as separate performance criteria.
However, this would result in four evaluation scores, which makes the evo-
lutionary search process more difficult. For this example, it was therefore
decided to combine these criteria into a single window performance score.
The saleable value and window performance evaluation criteria are in
conflict with one another. For a high saleable value, larger number of low
height blocks is preferred since it maximises the number of garden flats in
the ground floor. However, this results in closely packed blocks that obstruct
one another. For a high window performance, smaller number of tall blocks
is therefore preferred. The evolutionary process allows designers to explore
the trade-offs between such conflicting performance criteria.
The evolutionary process was executed on the Amazon EC2 cloud compu-
ting platform using Dexen, a distributed execution environment for popula-
tion based optimisation algorithms (Janssen et. al. 2011). Compute instances
were started with a total of 200 CPUs.
Figure 6: Final non-dominated Pareto set.
The population size was set to 200 and a simple asynchronous steady-state
evolutionary algorithm was used. Each generation, 50 individuals were ran-
domly selected from the population and ranked using multi-objective Pareto
ranking. The 2 individuals with the lowest rank were killed, and the 2 indi-
viduals with the highest rank (rank 1) were used as parents for reproduction.
Standard crossover and mutation operators for real-valued genotypes were
used, with a mutation probability of 0.02 and crossover probability of 0.9.
Reproduction between pairs of parents resulted in 2 new children, thereby
ensuring that the population size remained constant.
The final non-dominated Pareto set for the whole population contains a
range of design variants with differing trade-offs between saleable value and
window performance. The Pareto graph is shown in Figure 6. Three of the
design variants from this non-dominated set are shown in figure 7.
Figure 7: Design variants.
4. Conclusions
This paper has proposed a template and a set of techniques for the creation
of development and evaluation procedures for evolutionary design. The de-
velopment procedure generates a sparse skeletal model adhering to a variety
of constraints. For combinatorial constraints, decision chain encoding tech-
niques are used, and for geometric constraints, dynamics solver techniques
are used. Each evaluation procedure calculates an evaluation score for a spe-
cific performance criterion. The skeleton model is used in order to generate a
more detailed domain specific model, which is then used for analysis and
simulation. The resulting performance data is then condensed into a single
evaluation score.
The techniques used in the development and evaluation procedures can
be created by designers with limited programming skills using VDM soft-
ware. A demonstration has been presented where the template is used to cre-
ate development and evaluation procedures for a large and complex residen-
tial housing project. In the demonstration, the development and evaluation
procedures are defined using a VDM software called Sidefx Houdini, lever-
aging the procedural modelling tools and dynamic solvers that exist within
the software.
Bentley, P.J., editor: 1999, Evolutionary Design by Computers, Morgan Kaufmann Publish-
ers, San Francisco, CA.
Bentley, P.J. and Corne, D.W., editors: 2002, Creative Evolutionary Systems, Academic
Press, London, UK.
Caldas, L.: 2001, An evolution-based generative design system: using adaptation to shape ar-
chitectural form, Doctoral dissertation, Massachusetts Institute of Technology.
Draghi, J., and Wagner, G.: 2008, Evolution of evolvability in a development model, Theoreti-
cal Population Biology, 62, 301315.w
Eiben, A.E., and Smith J.E.: 2003, Introduction to evolutionary computing. In: Springer, Nat-
ural Computing Series, 1st edition.
Frazer, J.H.: 1995, An Evolutionary Architecture. AA Publications, London, UK
Janssen, P.H.T.: 2004, A design method and computational architecture for generating and
evolving building designs. Doctoral Dissertation, School of Design, Hong Kong Poly-
technic University.
Janssen, P.H.T.; Chen, K.W., and Basol, C.: 2011, Evolutionary development design for non-
programmers, In Proceedings of 29th eCAADe Conference, Ljubljana (Slovenia), 245-252
Janssen, P.H.T. and Kaushik, V.: 2012, Iterative Design Simulation: Exploring trade-offs be-
tween speed and accuracy. In Proceedings of 30th eCAADe Conference, Czech Technical
University, Czech Republic.
Janssen, P. H. T. and Kaushik, V.: 2013a, Skeletal modelling a developmental template for
evolutionary design. In Proceedings of the 18th International Conference on Computer-
Aided Architectural Design Research in Asia (CAADRIA), Singapore, 15-18 May 2013,
Janssen, P. H. T. and Kaushik, V.: 2013b, Decision chain encoding: Evolutionary design op-
timization with complex constraints. In Proceedings of the 2nd EvoMUSART Conference,
Vienna, Austria, 3-5 April 2013, 157167.
Kumar, S., and Bentley, P.J..: 1999, Three ways to grow designs: a comparison of embryoge-
nies for an evolutionary design problem, In Proceedings of the Genetic and Evolutionary
Computation Conference ‘99, 3543.
Mahner, M. and Kary, M.: 1997, What exactly are genomes, genotypes and phenotypes? And
what about phenomes? J. Theor. Biol, 186, 5563.
Michalewicz, Z.: 1996, Genetic Algorithms + Data Structures = Evolution Programs, Berlin
Heidelberg: Springer-Verlag, 3rd edition (First edition 1992).
... In this regard, novel and efficient design methods and techniques are being explored for improving the overall performance of HDB precinct designs with respect to various liveability criteria. Over the past decades, a number of studies have been conducted to integrate techniques such as parametric design and computational optimization into the design for HDB precincts Janssen & Kaushik, 2013;von Richthofen et al., 2018). These studies demonstrate the potential of computational optimization and performance-based design in assisting architects in their design of HDB precincts. ...
... Using this approach, buildings are typically placed over a predefined grid, thus resulting in a barracks-like building layout configuration von Richthofen et al., 2018). In contrast, the under-constrained case results in a lack of designs that are deemed feasible, as buildings are typically placed arbitrarily within a predefined boundary Janssen & Kaushik, 2013). Both cases can ultimately defeat the purpose of applying the optimization algorithm in the first place. ...
Conference Paper
Full-text available
This paper presents a ‘skeletal’ parametric schema to generate residential building layout configurations for performance-based design optimization. The schema generates residential building layout configurations using a set of ‘skeletal’ lines that are created based on various design elements and coincident with factors such as walkways, spacing, and setback requirements. As such, the schema is able to generate diverse and legitimate design alternatives. With the proposed parametric schema, a case-study optimization is carried out for a Singapore Housing Development Board (HDB) project. The case study considers a set of performance criteria and produces results with higher practical referential value. The case study demonstrates that the optimization with the parametric schema can improve the overall quality of the design and provide designers with various design options.
... An effective study of multi-criteria optimisation using EO's includes Janssen and Kaushik's [18] exploration of an approach that evolved populations of housing design variants. A template was used to create development and evaluation procedures for a large residential housing project. ...
Conference Paper
Full-text available
Performance-based design approaches typically use iterative simulation as a way of exploring design variants. For such approaches, the speed of execution of the simulations is critical to enabling a fluid and interactive design process. This research proposes an iterative simulation design method where simulations are configured to run in two modes: in fast mode, simulations produce less accurate results but due to their speed can be applied successfully within an iterative refinement process; in slow mode, the simulations produce more accurate results that can be used to verify the performance improvements achieved using the iterative refinement process. A case study is presented where the proposed method is used to explore performance improvements related to levels of incident illuminance and incident irradiance on windows
Conference Paper
Full-text available
A novel encoding technique is presented that allows constraints to be easily handled in an intuitive way. The proposed encoding technique structures the genotype-phenotype mapping process as a sequential chain of decision points, where each decision point consists of a choice between alternative options. In order to demonstrate the feasibility of the decision chain encoding technique, a case-study is presented for the evolutionary optimization of the architectural design for a large residential building.
Conference Paper
Full-text available
This paper explores the use of growth processes, or embryogenies, to map genotypes to phenotypes within evolutionary systems. Following a summary of the significant features of embryogenies, the three main types of embryogenies in Evolutionary Computation are then identified and explained: external, explicit and implicit. An experimental comparison between these three different embryogenies and an evolutionary algorithm with no embryogeny is performed. The problem set to the four evolutionary systems is to evolve tessellating tiles. In order to assess the scalability of the embryogenies, the problem is increased in difficulty by enlarging the size of tiles to be evolved. The results are surprising, with the implicit embryogeny outperforming all other techniques by showing no significant increase in the size of the genotypes or decrease in accuracy of evolution as the scale of the problem is increased.
Full-text available
"October 2004" Thesis (Ph.D.)--The Hong Kong Polytechnic University, 2005. Includes bibliographical references.
This dissertation dwells in the interstitial spaces between the fields of architecture, environmental design and computation. It introduces a Generative Design System that draws on evolutionary concepts to incorporate adaptation paradigms into the architectural design process. The initial aim of the project focused on helping architects improving the environmental performance of buildings, but the final conclusions of the thesis transcend this realm to question the process of incorporating computational generative systems in the broader context of architectural design. The Generative System [GS] uses a Genetic Algorithm as the search and optimization engine. The evaluation of solutions in terms of environmental performance is done using DOE2.1E. The GS is first tested within a restricted domain, where the optimal solution is previously known, to allow for the evaluation of the system's performance in locating high quality solutions. Results are very satisfactory and provide confidence to extend the GS to complex building layouts. Comparative studies using other heuristic search procedures like Simulated Annealing are also performed. The GS is then applied to an existing building by Alvaro Siza, to study the system's behavior in a complex architectural domain, and to assess its capability for encoding language constraints, so that solutions generated may be within certain design intentions. An extension to multicriteria problems is presented, using a Pareto-based method.
In "An Evolutionary Architecture", John Frazer presents an overview of his work for the past 30 years. Attempting to develop a theoretical basis for architecture using analogies with nature's processes of evolution and morphogenesis. Frazer's vision of the future of architecture is to construct organic buildings. Thermodynamically open systems which are more environmentally aware and sustainable physically, sociologically and economically. The range of topics which Frazer discusses is a good illustration of the breadth and depth of the evolutionary design problem. Environmental Modelling One of the first topics dealt with is the importance of environmental modelling within the design process. Frazer shows how environmental modelling is often misused or misinterpreted by architects with particular reference to solar modelling. From the discussion given it would seem that simplifications of the environmental models is the prime culprit resulting in misinterpretation and misuse. The simplifications are understandable given the amount of information needed for accurate modelling. By simplifying the model of the environmental conditions the architect is able to make informed judgments within reasonable amounts of time and effort. Unfortunately the simplications result in errors which compound and cause the resulting structures to fall short of their anticipated performance. Frazer obviously believes that the computer can be a great aid in the harnessing of environmental modelling data, providing that the same simplifying assumptions are not made and that better models and interfaces are possible. Physical Modelling Physical modelling has played an important role in Frazer's research. Leading to the construction of several novel machine readable interactive models, ranging from lego-like building blocks to beermat cellular automata and wall partitioning systems. Ultimately this line of research has led to the Universal Constructor and the Universal Interactor. The Universal Constructor The Universal Constructor features on the cover of the book. It consists of a base plug-board, called the "landscape", on top of which "smart" blocks, or cells, can be stacked vertically. The cells are individually identified and can communicate with neighbours above and below. Cells communicate with users through a bank of LEDs displaying the current state of the cell. The whole structure is machine readable and so can be interpreted by a computer. The computer can interpret the states of the cells as either colour or geometrical transformations allowing a wide range of possible interpretations. The user interacts with the computer display through direct manipulation of the cells. The computer can communicate and even direct the actions of the user through feedback with the cells to display various states. The direct manipulation of the cells encourages experimentation by the user and demonstrates basic concepts of the system. The Universal Interactor The Universal Interactor is a whole series of experimental projects investigating novel input and output devices. All of the devices speak a common binary language and so can communicate through a mediating central hub. The result is that input, from say a body-suit, can be used to drive the out of a sound system or vice versa. The Universal Interactor opens up many possibilities for expression when using a CAD system that may at first seem very strange.However, some of these feedback systems may prove superior in the hands of skilled technicians than more standard devices. Imagine how a musician might be able to devise structures by playing melodies which express the character. Of course the interpretation of input in this form poses a difficult problem which will take a great deal of research to achieve. The Universal Interactor has been used to provide environmental feedback to affect the development of evolving genetic codes. The feedback given by the Universal Interactor has been used to guide selection of individuals from a population. Adaptive Computing Frazer completes his introduction to the range of tools used in his research by giving a brief tour of adaptive computing techniques. Covering topics including cellular automata, genetic algorithms, classifier systems and artificial evolution. Cellular Automata As previously mentioned Frazer has done some work using cellular automata in both physical and simulated environments. Frazer discusses how surprisingly complex behaviour can result from the simple local rules executed by cellular automata. Cellular automata are also capable of computation, in fact able to perform any computation possible by a finite state machine. Note that this does not mean that cellular automata are capable of any general computation as this would require the construction of a Turing machine which is beyond the capabilities of a finite state machine. Genetic Algorithms Genetic algorithms were first presented by Holland and since have become a important tool for many researchers in various areas.Originally developed for problem-solving and optimization problems with clearly stated criteria and goals. Frazer fails to mention one of the most important differences between genetic algorithms and other adaptive problem-solving techniques, ie. neural networks. Genetic algorithms have the advantage that criteria can be clearly stated and controlled within the fitness function. The learning by example which neural networks rely upon does not afford this level of control over what is to be learned. Classifier Systems Holland went on to develop genetic algorithms into classifier systems. Classifier systems are more focussed upon the problem of learning appropriate responses to stimuli, than searching for solutions to problems. Classifier systems receive information from the environment and respond according to rules, or classifiers. Successful classifiers are rewarded, creating a reinforcement learning environment. Obviously, the mapping between classifier systems and the cybernetic view of organisms sensing, processing and responding to environmental stimuli is strong. It would seem that a central process similar to a classifier system would be appropriate at the core of an organic building. Learning appropriate responses to environmental conditions over time. Artificial Evolution Artificial evolution traces it's roots back to the Biomorph program which was described by Dawkins in his book "The Blind Watchmaker". Essentially, artificial evolution requires that a user supplements the standard fitness function in genetic algorithms to guide evolution. The user may provide selection pressures which are unquantifiable in a stated problem and thus provide a means for dealing ill-defined criteria. Frazer notes that solving problems with ill-defined criteria using artificial evolution seriously limits the scope of problems that can be tackled. The reliance upon user interaction in artificial evolution reduces the practical size of populations and the duration of evolutionary runs. Coding Schemes Frazer goes on to discuss the encoding of architectural designs and their subsequent evolution. Introducing two major systems, the Reptile system and the Universal State Space Modeller. Blueprint vs. Recipe Frazer points out the inadequacies of using standard "blueprint" design techniques in developing organic structures. Using a "recipe" to describe the process of constructing a building is presented as an alternative. Recipes for construction are discussed with reference to the analogous process description given by DNA to construct an organism. The Reptile System The Reptile System is an ingenious construction set capable of producing a wide range of structures using just two simple components. Frazer saw the advantages of this system for rule-based and evolutionary systems in the compactness of structure descriptions. Compactness was essential for the early computational work when computer memory and storage space was scarce. However, compact representations such as those described form very rugged fitness landscapes which are not well suited to evolutionary search techniques. Structures are created from an initial "seed" or minimal construction, for example a compact spherical structure. The seed is then manipulated using a series of processes or transformations, for example stretching, shearing or bending. The structure would grow according to the transformations applied to it. Obviously, the transformations could be a predetermined sequence of actions which would always yield the same final structure given the same initial seed. Alternatively, the series of transformations applied could be environmentally sensitive resulting in forms which were also sensitive to their location. The idea of taking a geometrical form as a seed and transforming it using a series of processes to create complex structures is similar in many ways to the early work of Latham creating large morphological charts. Latham went on to develop his ideas into the "Mutator" system which he used to create organic artworks. Generalising the Reptile System Frazer has proposed a generalised version of the Reptile System to tackle more realistic building problems. Generating the seed or minimal configuration from design requirements automatically. From this starting point (or set of starting points) solutions could be evolved using artificial evolution. Quantifiable and specific aspects of the design brief define the formal criteria which are used as a standard fitness function. Non-quantifiable criteria, including aesthetic judgments, are evaluated by the user. The proposed system would be able to learn successful strategies for satisfying both formal and user criteria. In doing so the system would become a personalised tool of the designer. A personal assistant which would be able to anticipate aesthetic judgements and other criteria by employing previously successful strategies. Ultimately, this is a similar concept to Negroponte's "Architecture Machine" which he proposed would be computer system so personalised so as to be almost unusable by other people. The Universal State Space Modeller The Universal State Space Modeller is the basis of Frazer's current work. It is a system which can be used to model any structure, hence the universal claim in it's title. The datastructure underlying the modeller is a state space of scaleless logical points, called motes. Motes are arranged in a close-packing sphere arrangement, which makes each one equidistant from it's twelve neighbours. Any point can be broken down into a self-similar tetrahedral structure of logical points. Giving the state space a fractal nature which allows modelling at many different levels at once. Each mote can be thought of as analogous to a cell in a biological organism. Every mote carries a copy of the architectural genetic code in the same way that each cell within a organism carries a copy of it's DNA. The genetic code of a mote is stored as a sequence of binary "morons" which are grouped together into spatial configurations which are interpreted as the state of the mote. The developmental process begins with a seed. The seed develops through cellular duplication according to the rules of the genetic code. In the beginning the seed develops mainly in response to the internal genetic code, but as the development progresses the environment plays a greater role. Cells communicate by passing messages to their immediate twelve neighbours. However, it can send messages directed at remote cells, without knowledge of it's spatial relationship. During the development cells take on specialised functions, including environmental sensors or producers of raw materials. The resulting system is process driven, without presupposing the existence of a construction set to use. The datastructure can be interpreted in many ways to derive various phenotypes. The resulting structure is a by-product of the cellular activity during development and in response to the environment. As such the resulting structures have much in common with living organisms which are also the emergent result or by-product of local cellular activity. Primordial Architectural Soups To conclude, Frazer presents some of the most recent work done, evolving fundamental structures using limited raw materials, an initial seed and massive feedback. Frazer proposes to go further and do away with the need for initial seed and start with a primordial soup of basic architectural concepts. The research is attempting to evolve the starting conditions and evolutionary processes without any preconditions. Is there enough time to evolve a complex system from the basic building blocks which Frazer proposes? The computational complexity of the task being embarked upon is not discussed. There is an implicit assumption that the "superb tactics" of natural selection are enough to cut through the complexity of the task. However, Kauffman has shown how self-organisation plays a major role in the early development of replicating systems which we may call alive. Natural selection requires a solid basis upon which it can act. Is the primordial soup which Frazer proposes of the correct constitution to support self-organisation? Kauffman suggests that one of the most important attributes of a primordial soup to be capable of self-organisation is the need for a complex network of catalysts and the controlling mechanisms to stop the reactions from going supracritical. Can such a network be provided of primitive architectural concepts? What does it mean to have a catalyst in this domain? Conclusion Frazer shows some interesting work both in the areas of evolutionary design and self-organising systems. It is obvious from his work that he sympathizes with the opinions put forward by Kauffman that the order found in living organisms comes from both external evolutionary pressure and internal self-organisation. His final remarks underly this by paraphrasing the words of Kauffman, that life is always to found on the edge of chaos. By the "edge of chaos" Kauffman is referring to the area within the ordered regime of a system close to the "phase transition" to chaotic behaviour. Unfortunately, Frazer does not demonstrate that the systems he has presented have the necessary qualities to derive useful order at the edge of chaos. He does not demonstrate, as Kauffman does repeatedly, that there exists a "phase transition" between ordered and chaotic regimes of his systems. He also does not make any studies of the relationship of useful forms generated by his work to phase transition regions of his systems should they exist. If we are to find an organic architecture, in more than name alone, it is surely to reside close to the phase transition of the construction system of which is it built. Only there, if we are to believe Kauffman, are we to find useful order together with environmentally sensitive and thermodynamically open systems which can approach the utility of living organisms.