I. Mitchell (UK)
Genetic Algorithms, Maximal Cliques, Graph Colouring.
This paper investigates the domain of two graph problems, maximal clique and graph colouring problems. The solution to such problems seeks an equal chromosome representation. Therefore given a graph, G(V,E), the same chromosome should yield two different results when applied to the associated problem domain. In this paper several graphs are used to test the algorithm and near-optimal results where yielded. However, it is not the results that are of interest but the concept of generic representations. GAs are used to find near-optimal results in many problems domains, however each problem domain is accompanied with its' specific chromosome representation. In this paper a proposal is made to extend this concept and present a generic representation i.e. a representation that can represent more than one problem domain. The benefits of generic representations are discussed and include the generation of genotype phenotype mappings, via genetic programming, to form better rule sets and can solutions to two problem domains be coupled? In this paper it is argued the discovery of couplings between two problem domains requires a generic representation. This research forms the foundation for further investigation into the issue of modeling enforced mutualism.
Important Links:
Go Back
