Keywords : Symbolic Regression

Improving Gene Expression Programming Method

Najla Akram Al-Saati; Nidhal Al-Assady

AL-Rafidain Journal of Computer Sciences and Mathematics, 2009, Volume 6, Issue 1, Pages 81-95
DOI: 10.33899/csmj.2009.163767

In this work the algorithm of Gene Expression Programming (GEP) is investigated thoroughly and the major deficiencies are pointed out. Multiple suggestions for enhancements are introduced in this research aiming at solving the major deficiencies that were investigated. These improvements produced higher success rates and avoid the malfunctioning situations found in GEP. These deficiencies or weak points include: choosing the best parameter settings, using only one linking function, gene flattening problem, illegal operations in genes and lack of function biasing. Improvements suggested the following enhancement features: the Multi-Population feature, the Emergency Mutation feature, and the feature of ComponentBiasing. Tests are carried out using two different symbolic regression problems.