2nd International ICST Conference on Bio-Inspired Models of Network, Information, and Computing Systems

Research Article

Introduction of a Sectioned Genetic Algorithm for Large Scale Problems

Download69 downloads
  • @INPROCEEDINGS{10.4108/ICST.BIONETICS2007.2404,
        author={Zacharias Detorakis and George Tambouratzis},
        title={Introduction of a Sectioned Genetic Algorithm for Large Scale Problems},
        proceedings={2nd International ICST Conference on Bio-Inspired Models of Network, Information, and Computing Systems},
        proceedings_a={BIONETICS},
        year={2008},
        month={8},
        keywords={Genetic algorithms  input space dimensionality  masks  parallel distributed algorithms  stemming},
        doi={10.4108/ICST.BIONETICS2007.2404}
    }
    
  • Zacharias Detorakis
    George Tambouratzis
    Year: 2008
    Introduction of a Sectioned Genetic Algorithm for Large Scale Problems
    BIONETICS
    ICST
    DOI: 10.4108/ICST.BIONETICS2007.2404
Zacharias Detorakis1,*, George Tambouratzis2,*
  • 1: Inst. for Language and Speech Processing 6 Artemidos & Epidavrou Str. Paradissos Amaroussiou, 151 25, Greece ++30 210-6875363
  • 2: Inst. for Language and Speech Processing 6 Artemidos & Epidavrou Str. Paradissos Amaroussiou, 151 25, Greece ++30 210-6875411
*Contact email: zdetor@ilsp.gr, giorg_t@ilsp.gr

Abstract

The sectioned genetic algorithm (hereafter denoted as sectioned GA), which is presented in this paper, represents a modification of the standard GA and deals with large scale problems (i.e. problems involving pattern spaces with high dimensionalities). Instead of increasing the size of the population searching the pattern space when the problem dimensionality increases, the sectioned GA approach divides each individual into smaller parts (sections) and subsequently applies the genetic operators on each of these parts. Results from the application of sectioned GA on the problem of automatic morphological analysis are also presented in this article. Morphological analysis is by nature a large scale problem since a great number of words need to be segmented into stems and suffixes. The proposed system improves the segmentation accuracy substantially in comparison to standard GA algorithms.