Improved differential evolution for dynamic optimization problems

This article reports improvements on DynDE, a approach to using Differential Evolution to solve dynamic optimization problems. Three improvements are suggested, namely favored populations, migrating individuals and a combination of these approaches. The effects of varying the change frequency, peak...

Full description

Saved in:
Bibliographic Details
Main Authors: du Plessis, M.C., Engelbrecht, A.P.
Format: Conference Proceeding
Language:English
Subjects:
Benchmark testing
Correlation
Evolution (biology)
Evolutionary computation
Heuristic algorithms
Optimization
Tracking
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:This article reports improvements on DynDE, a approach to using Differential Evolution to solve dynamic optimization problems. Three improvements are suggested, namely favored populations, migrating individuals and a combination of these approaches. The effects of varying the change frequency, peak widths and the number of dimensions of the dynamic environment are investigated. Experimental results are presented that indicate that the suggested approaches constitute considerable improvements on previous research.
ISSN:1089-778X
1941-0026
DOI:10.1109/CEC.2008.4630804