Journal of Applied and Physical Sciences
Details
Journal ISSN: 2414-3103
Article DOI: https://doi.org/10.20474/japs-2.2.4
Received: 15 April 2016
Accepted: 23 May 2016
Published: 23 June 2016
Download Article (PDF)
  • Two Enhanced Differential Evaluation Algorithms for expensive optimization


Muneer Maaroof Hasan, Dr. Oğuz Altun

Article first published online: 2016

Abstract

In this paper, the meta-heuristic algorithm which named Differential Evaluation (DE) has been improved. The improving made to increase the exploration rate and decrease the run time. Since DE needs too long time, when we implement it to solve computational expensive problems, we developed two different versions of DE named by Enhanced1 Differential Evaluation (E1DE) and Enhanced2 Differential Evaluation (E2DE). E1DE and E2DE were introduced to solve Computationally Expensive Optimization (CEO). Problems discussed and tested using all 15 test functions of the Special Session & Competition on Real-Parameter Single Objective Optimization (Expensive Case) at Congress on Evolutionary Computation 2015 (CEC-2015). The results show that the work significantly improved the basic DE in time by 54% and in results by 86%.