By Topic

Multiobjectivization via Helper-Objectives With the Tunable Objectives Problem

Sign In

Cookies must be enabled to login.After enabling cookies , please use refresh or reload or ctrl+f5 on the browser for the login options.

Formats Non-Member Member
$33 $13
Learn how you can qualify for the best price for this item!
Become an IEEE Member or Subscribe to
IEEE Xplore for exclusive pricing!
close button

puzzle piece

IEEE membership options for an individual and IEEE Xplore subscriptions for an organization offer the most affordable access to essential journal articles, conference papers, standards, eBooks, and eLearning courses.

Learn more about:

IEEE membership

IEEE Xplore subscriptions

2 Author(s)
Lochtefeld, D.F. ; 711th Human Performance Wing, Air Force Res. Lab., Wright-Patterson AFB, OH, USA ; Ciarallo, F.W.

Multiobjectivization, the optimization of a single-objective problem by adding objectives, has recently received interest by researchers. Studying multiobjectivization on an abstract problem can assist in understanding the fundamental drivers of the improvements in performance that multiobjectivization achieves in some situations. Previously created abstract problems do not appear to provide the modeling power needed to study the benefits of this new family of optimization techniques. The tunable objectives problem (TOP) model is introduced to help demonstrate how problem features such as objective-convolution, multiple layer epistasis, the presence of local optima, and layered problem structure are related to the performance of multiobjectivization via helper objectives. Experiments using TOP demonstrate how multiobjectivization via helpers improves the signal-to-noise in a genetic algorithm and identifies several general problem difficulties that, when present, are likely to increase the need for multiobjectivization.

Published in:

Evolutionary Computation, IEEE Transactions on  (Volume:16 ,  Issue: 3 )