Property | Value |
Name | NCL-TR-2010007 |
Description |
Linkage Identification Based on Decision Tree Learning
Abstract: Genetic algorithms and their descendant methods have been deemed robust, effective, and practical for the past decades. In order to enhance the features and capabilities of genetic algorithms, tremendous effort has been invested within the research community of evolutionary computation. One of the major development trends to improve genetic algorithms is trying to extract and exploit the relationship among decision variables, such as estimation of distribution algorithms and perturbation-based methods. In this study, we attempt to investigate the integration of perturbation-based method, inductive linkage identification (ILI), and decision tree learning algorithms for detecting general problem structures. Experiments on circular problems structures composed of order-4 and order-5 trap functions are conducted, as well as the experiments on the cardinality of variables. Our experiments results indicates that this approach requires a population size growing logarithmically and is insensitive to the problem structure consisting of similar sub-structures as long as the overall problem size is identical and those variables are binary. Experiments also show that different adopted decision tree learning algorithms will demand different requirements when the variables are extended to higher cardinalities. |
Filename | NCL-TR-2010007.pdf |
Filesize | 541.39 kB |
Filetype | pdf (Mime Type: application/pdf) |
Creator | ypchen |
Created On: | 09/01/2010 17:16 |
Viewers | Everybody |
Maintained by | Editor |
Hits | 2475 Hits |
Last updated on | 12/09/2010 12:14 |
Homepage |