Generative Representations for Evolving Families of Designs

Hornby, Gregory S.

doi:10.1007/3-540-45110-2_61

Cited by 40 publications

(43 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Hornby looked at a wide variety of representation schemes for generative design [4], including "virtual creatures" generated using a variant of L-systems. Hornby and others [2] have observed that the generative reuse of parameterised elements in encoded designs improves the ability of an evolutionary algorithm to search large design spaces.…”

Section: Related Workmentioning

confidence: 99%

Balancing Act: Variation and Utility in Evolutionary Art

McCormack

2014

Evolutionary and Biologically Inspired Music, Sound, Art and Design

View full text Add to dashboard Cite

Abstract. Evolutionary Art typically involves a tradeoff between the size and flexibility of genotype space and its mapping to an expressive phenotype space. Ideally we would like a genotypic representation that is terse but expressive, that is, we want to maximise the useful variations the genotype is capable of expressing in phenotype space. Terseness is necessary to minimise the size of the overall search space, and expressiveness can be loosely interpreted as phenotypes that are useful (of high fitness) and diverse (in feature space). In this paper I describe a system that attempts to maximise this ratio between terseness and expressiveness. The system uses a binary string up to any maximum length as the genotype. The genotype string is interpreted as building instructions for a graph, similar to the cellular programming techniques used to evolve artificial neural networks. The graph is then interpreted as a formbuilding automaton that can construct animated 3-dimensional forms of arbitrary complexity. In the test case the requirement for expressiveness is that the resultant form must have recognisable biomorphic properties and that every possible genotype must fulfil this condition. After much experimentation, a number of constraints in the mapping technique were devised to satisfy this condition. These include a special set of geometric building operators that take into account morphological properties of the generated form. These methods were used in the evolutionary artwork "Codeform", developed for the Ars Electronica museum. The work generated evolved virtual creatures based on genomes acquired from the QR codes on museum visitor's entry tickets.

show abstract

Section: Related Workmentioning

confidence: 99%

Balancing Act: Variation and Utility in Evolutionary Art

McCormack

2014

Evolutionary and Biologically Inspired Music, Sound, Art and Design

View full text Add to dashboard Cite

show abstract

“…Boers and Sprinkhuizen-Kuyper [15] used a string L-system to also evolve ANNs by interpreting the final string that results from a given number of rewrites as a graph. The grammar of GENRE [16], an evolutionary design framework based on a parametric L-system, is evolved by a simple EA with specialized operators. Strings are rewritten and then translated into designs, with successful application to table designs, neural networks, and robot controllers.…”

Section: Grammatical Developmentmentioning

confidence: 99%

Graph design by graph grammar evolution

Luerssen

Powers

2007

2007 IEEE Congress on Evolutionary Computation

View full text Add to dashboard Cite

Abstract-Determining the optimal topology of a graph is pertinent to many domains, as graphs can be used to model a variety of systems. Evolutionary algorithms constitute a popular optimization method, but scalability is a concern with larger graph designs. Generative representation schemes, often inspired by biological development, seek to address this by facilitating the discovery and reuse of design dependencies and allowing for adaptable exploration strategies. We present a novel developmental method for optimizing graphs that is based on the notion of directly evolving a hypergraph grammar from which a population of graphs can be derived. A multi-objective design system is established and evaluated on problems from three domains: symbolic regression, circuit design, and neural control. The observed performance compares favorably with existing methods, and extensive reuse of subgraphs contributes to the efficient representation of solutions. Constraints can also be placed on the type of explored graph spaces, ranging from tree to pseudograph. We show that more compact solutions are attainable in less constrained spaces, although convergence typically improves with more constrained designs.

show abstract

“…Some researchers tried to systematize and compare arti cial genetic encodings in a single phenetic model [28,6,20,15]. This allows to test their performance in the same environment, and thus focus on the characteristics of the encodings (features like their complexity, constraints, support for modularity, body symmetry, compression, redundancy, etc.).…”

Section: Genetics In Arti Cial Life Systemsmentioning

confidence: 99%

“…The freedom in designing and using various genetic encodings allows researchers to compare them and study their properties in a single simulation model [28,6,15,20]. If many genetic encodings exist in a single simulation model, then each of them must be capable of expressing phenotypes in that model.…”

Section: Introductionmentioning

confidence: 99%

Genetic mappings in artificial genomes

Komosiński

2004

Theory in Biosciences

View full text Add to dashboard Cite

This paper concerns processing of genomes of arti cial (computer-simulated) organisms. Of special interest is the process of translation of genotypes into phenotypes, and utilizing the mapping information obtained during such translation. If there exists more than one genetic encoding in a single arti cial life model, then the translation may also occur between di erent encodings. The obtained mapping information allows to present genes-phenes relationships visually and interactively to a person, in order to increase understanding of the genotype-to-phenotype translation process and genetic encoding properties. As the mapping associates parts of the source sequence with the translated destination, it may be also used to trace genes, phenes, and their relationships during simulated evolution.A mappings composition procedure is formally described, and a simple method of visual mapping presentation is established. Finally, advanced visualizations of genephene relationships are demonstrated as practical examples of introduced techniques. These visualizations concern genotypes expressed in various encodings, including an encoding which exhibits polygenic and pleiotropic properties.

show abstract

Generative Representations for Evolving Families of Designs

Cited by 40 publications

References 11 publications

Balancing Act: Variation and Utility in Evolutionary Art

Balancing Act: Variation and Utility in Evolutionary Art

Graph design by graph grammar evolution

Genetic mappings in artificial genomes

Contact Info

Product

Resources

About