This paper analyzes the genetic operations of a new evolution mechanism proposed by us for improving the capability to deal with graph-form solutions in the real world of genetic algorithms based on the theories of GAs and GPs. A prototype of graph evolution with genetic operations is implemented and applied to some graph-related systems with the Irish-student classification data. Evaluation between conventional optimization mechanisms and graph evolution theory is also made for proving the advantage of using graph evolution. Be notable is the graph evolution theory proposed in this paper can cover most applications of GAs and GPs.
|Number of pages||6|
|Publication status||Published - 2003|
|Event||International Joint Conference on Neural Networks 2003 - Portland, OR, United States|
Duration: 20 Jul. 2003 → 24 Jul. 2003
|Conference||International Joint Conference on Neural Networks 2003|
|Period||20/07/03 → 24/07/03|