We proposed and studied from the point of view of formal languages the novel formalism of string-regulated graph grammars, as an abstraction ofbiological development. Founded on this formalism, I introduced and tested evolutionary developmental models based on genetic regulation andtensegrity mechanisms. In the area of evolutionary computation, the results demonstrate the benefits of the models, compared to traditional direct encoding approaches, for the problems of form-finding and behavior-finding. Such benefits include high diversity, regularity, organicity, generalization, and tensegrity modularity; properties that are highly valuable in an engineering context.

Additionally, we have employed these computational models to gain further insight into the evolutionary origins of novel features and diversity, responsible for the extraordinary biodiversity of our planet. Supportingthe latest biological hypothesis, the findings show that the sole variation in the regulation of gene expression, rather than the variation in the structural genes, is indeed a sufficient condition for the evolutionary emergence of novelty and diversity.

Finally, since natural evolution is a great example of a huge creative process, the presented evolutionary developmental models are also especially suitable for computational creative applications. More precisely, we have applied the models in a novel methodology for the automatic animation of characters for computer graphics.

In summary, I have pursued a truly interdisciplinary research path, with contributions to the fields of formal languages, evolutionary computation, evolutionary biology, developmental biology, artificial life, engineering design optimization, and computer animation.

My thesis is available at this link.

I have programmed the models and simulators used in this work in Matlab. You can freely download, under a GPL license, the source code of a tensegrity physics simulator based on springs (MatSprings) and of an evolutionary developmental simulator (MatEvoDevo).Notice that, due to the high computational cost required running the physical simulations, the use of a computer cluster is advised. The software provided is already adapted to benefit from running in a computer cluster or a multi-core computer.