AncDE with gaussian distribution for numerical optimization problem

Siti Khadijah Mohd Salleh (Universiti Teknikal Malaysia Melaka)
Siti Azirah Asmai (Universiti Teknikal Malaysia Melaka)
Zuraida Abal Abas (Universiti Teknikal Malaysia Melaka)
Abdul Samad Shibghatullah (UCSI University Kuala Lumpur)
Diarmuid O'Donoghue (Maynooth University)

International Journal of Advances in Intelligent Informatics
Vol 5, No 1 (2019): March 2019

This work is introducing an enhanced Differential Evolution (DE) called AncDE. This proposed algorithm is using an additional population from the current generation and located it as ancestor. There are two parameter controllers to manage the selection of ancestor vector; aup for selection frequency and arp for age of selection. In this work we were applying Gaussian distribution on aup and we tested it on CEC 2015 Numerical Optimization Problem. Standard Differential Evolution will act as the benchmark. The result shows that AncDE with Gaussian approach has produced better result than standard DE.