An investigation on mixing heterogeneous differential evolution variants in a distributed framework
by S. Thangavelu; C. Shunmuga Velayutham
International Journal of Bio-Inspired Computation (IJBIC), Vol. 7, No. 5, 2015

Abstract: This paper attempts a preliminary investigation to gain insight about the cooperative dynamics of mixing the four classical differential evolution (DE) variants viz. DE/rand/1/bin, DE/best/1/bin, DE/rand/2/bin and DE/best/2/bin in an island-based distributed framework. The exhaustive combinations of the above said four DE variants in an island size of 4, resulting in 35 distributed DE variants, have all been implemented and tested on 14 unconstrained test functions with diverse features grouped by their modality and decomposability. Simulation results show that the rand-best variants' mixing, display a better cooperative characteristics than rand-rand and best-best variants' mixing. This insight motivated for further investigations on mixing DE/rand-to-best/1/bin (a variant which intrinsically employs rand and best strategies) with DE/rand/1/bin and DE/best/1/bin in the distributed framework. Simulation results reiterated the observations about the cooperative characteristics of rand-best variants' combinations with the latter mixing showing still better cooperative characteristics both in terms of probability of convergence and convergence rate.

Online publication date: Wed, 07-Oct-2015

The full text of this article is only available to individual subscribers or to users at subscribing institutions.

 
Existing subscribers:
Go to Inderscience Online Journals to access the Full Text of this article.

Pay per view:
If you are not a subscriber and you just want to read the full contents of this article, buy online access here.

Complimentary Subscribers, Editors or Members of the Editorial Board of the International Journal of Bio-Inspired Computation (IJBIC):
Login with your Inderscience username and password:

    Username:        Password:         

Forgotten your password?


Want to subscribe?
A subscription gives you complete access to all articles in the current issue, as well as to all articles in the previous three years (where applicable). See our Orders page to subscribe.

If you still need assistance, please email subs@inderscience.com