E-PROCEEDING OF THE 5TH INTERNATIONAL CONFERENCE ON ARTIFICIAL INTELLIGENCE, COMPUTER SCIENCE & INFORMATION TECHNOLOGY 2017 100 E-PROCEEDING OF ARTIFICIAL INTELLIGENCE, COMPUTER SCIENCE & INFORMATION TECHNOLOGY (e-ISBN: 978-967-0792-19-4). 31st July 2017& 1st August 2017 - Venue: Bayview Hotel, Pulau Pinang, Malaysia. Organized by http://worldconferences.net/home MAGNITUDE OF DIFFERENCE (MDV) BETWEEN TARGET VECTOR AND TRIAL VECTOR IN ANCDE AND STANDARD DIFFERENTIAL EVOLUTION Siti Khadijah Mohd Salleh Dept. Of Info. Tech. and Comm Politeknik Ungku Omar Perak, Malaysia sitikhadijah@puo.edu.my Diarmuid O’Donoghue Dept. Of Computer Science Maynooth University Co. Kildare, Ireland Diarmuid.ODonoghue@nuim.ie Abdul Samad Shibghatullah Dept. of Info and Comm. Tech UTEM Melaka, Malaysia samad@utem.edu.my Abstract This paper describes an ancestral extension to the standard Differential Evolution algorithm called AncDE and attempts to explain why its ancestral cache helps improve performance. We compute the magnitude difference between the target vector and the trial vector for each generation for DE and AncDE. Results for 6 problems are used in this study. Our findings indicate that ancestors can help overcome some of the local variation in solutions quality and improve solution quality by improving population diversity. Keywords: Ancestor template, Differential Evolution, magnitude difference vector. I. INTRODUCTION Differential Evolution (DE) has become extremely popular because of its efficiency and simple implementation and has been shown to be one of the most reliable algorithms in dealing with optimization problems [1]. In standard DE, three distinct target vectors r1, r2, r3 are randomly selected from initial population G = {x1, x2,...,x|G|}. In mutation stage, F is a mutation factor in [0, 2] that controls the extension of differential variation (Xr2,G – Xr3,G) then added with the third target vector. Crossover phase produces a trial vector:  ,, ={  ,, , if (  [0,1] ≤ ) ( =  )  ,, , ℎ ,  = 1,2 … 