Evolvable Hardware
Computer Architecture and Design Group,  IDI, NTNU

Shrinking the Genotype

The scalability problem is due to the resource-greedy nature of evolutionary techniques. Generally in EHW, a one-to-one mapping has
been chosen for the genotype-phenotype transition.  This means that the genotype needs to include all the information required for the
phenotype i.e. configuration data to program the device. The larger the circuit to be evolved, the more logic and routing information
required. As the complexity of the genotype increases, so increases the computational and storage requirements.

A solution to this problem is to shrink the genotype in some way. This may be stated as the genotype challenge, finding new forms of
representation for evolving complex circuits. In this work we are interested in investigating ways in which development may be combined
with genetic algorithms so as to enable smaller genotypes to evolve complex circuit designs.

[top]

Participants

Faculty
Pauline Haddow, Assoc. Professor, Department of Computer and Information Science, NTNU

Students
Gunnar Tufte, PhD student; Department of Computer and Information Science, NTNU
Piet van Remortel, PhD student, NTNU and Free University of Brussels

Current Projects

[top]
L-systems for EHW
 

Relevant Publications

"Shrinking the Genotype: L-systems for EHW?
by Pauline C Haddow, Gunnar Tufte and Piet van Remortel, submitted to ICES'2001

"Bridging the Genotype-Phenotype Mapping for Digital FPGAs"
by Pauline C Haddow and Gunnar Tufte, to appear in EH'01