Details for NCL-TR-2009009
PropertyValue
NameNCL-TR-2009009
Description
Inductive Linkage Identification on Building Blocks of Different Sizes and Types
Chen, Ying-ping, Chuang, Chung-Yao, & Huang, Yuan-Wei
Abstract: The goal of linkage identification is to obtain the dependencies among decision variables. Such information or knowledge can be applied to the designs of crossover operators and/or the encoding schemes in genetic and evolutionary methods. Thus, promising sub-solutions to the problem will be less probably disrupted and successful convergences may more likely to be achieved. In our previous studies, a linkage identification technique, called Inductive Linkage Identification (ILI), was proposed. This method was established upon the mechanism of perturbation and the idea of decision tree learning. By constructing a decision tree according to decision variables and resulting fitness difference values, the interdependent variables will be determined by the decision tree learning algorithm. In this paper, we aim to acquire more understandings on the characteristics of ILI, especially its behavior under problems composed of different-sized and different-typed building blocks. Experiments showed that ILI can efficiently handle building block of different sizes and is insensitive to building block types. Our experimental observations indicate the flexibility and the applicability of ILI on various elementary building block types that are commonly adopted in experiments.
FilenameNCL-TR-2009009.pdf
Filesize260.08 kB
Filetypepdf (Mime Type: application/pdf)
Creatorypchen
Created On: 11/29/2009 20:49
ViewersEverybody
Maintained byEditor
Hits2567 Hits
Last updated on 02/23/2011 18:34
Homepage