Improving differential evolution through a unified approach

Nikhil Padhye (), Piyush Bhardawaj () and Kalyanmoy Deb ()

Journal of Global Optimization, 2013, vol. 55, issue 4, 799 pages

Abstract: Only a few attempts in past have been made in adopting a unified outlook towards different paradigms in evolutionary computation (EC). The underlying motivation of these studies was aimed at gaining better understanding of evolutionary methods, both at the level of theory as well as application, in order to design efficient evolutionary algorithms for solving wide-range of complex problems. However, the past descriptions have either been too general or sometimes abstract in issuing a clear direction for improving an evolutionary paradigm for a task-specific. This paper recollects the ‘Unified Theory of Evolutionary Computation’ from past and investigates four steps—Initialization, Selection, Generation and Replacement, which are sufficient to describe traditional forms of Evolutionary Optimization Systems such as Genetic Algorithms, Evolutionary Strategies, Evolutionary Programming, Particle Swarm Optimization and differential evolution (DE). Then, a relatively new evolutionary paradigm, DE, is chosen and studied for its performance on a set of unimodal problems. Discovering DEs inability as an efficient solver, DE is reviewed under ‘Unified Framework’ and functional requirements of each step are evaluated. Targeted towards enhancing the DE’s performance, several modifications are proposed through borrowing of operations from a benchmark solver G3-PCX. Success of this exercise is demonstrated in a step-by-step fashion via simulation results. The Unified Approach is highly helpful in understanding the role and re-modeling of DE steps in order to efficiently solve unimodal problems. In an avalanching-age of new methods in EC, this study outlines a direction for advancing EC methods by undertaking a collective outlook and an approach of concept-sharing. Copyright Springer Science+Business Media, LLC. 2013

Keywords: Evolutionary computation; Optimization; Genetic algorithms; Differential evolution; Computational complexity (search for similar items in EconPapers)
Date: 2013
References: View complete reference list from CitEc
Citations: View citations in EconPapers (5)

Downloads: (external link)
http://hdl.handle.net/10.1007/s10898-012-9897-0 (text/html)
Access to full text is restricted to subscribers.

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:spr:jglopt:v:55:y:2013:i:4:p:771-799

Ordering information: This journal article can be ordered from
http://www.springer. ... search/journal/10898

DOI: 10.1007/s10898-012-9897-0

Access Statistics for this article

Journal of Global Optimization is currently edited by Sergiy Butenko

More articles in Journal of Global Optimization from Springer
Bibliographic data for series maintained by Sonal Shukla () and Springer Nature Abstracting and Indexing ().