Generic Representation of Solid-Object Geometry for Genetic Search

This paper examines the first stage of using genetic algorithms in design - how to encode designs as chromosomes. Generic representations capable of describing the geometry of any solid object to eliminate the unnecessary duplication of work at this stage are explored. A suitable low-parameter spatial partitioning representation is then proposed, using partitions consisting of six-sided polyhedra intersected by planes. Finally, the coding of the representation is examined, with the conclusion that allelic coding with control
genes and user-definable fixed-value alleles
would provide the most flexible system.