Ocean Feature Recognition Using Genetic Algorithms with Fuzzy Fitness Functions (GA/F3)

Abstract

Genetic algorithms are a problem solving method requiring domain- specific knowledge that is often heuristic. Candidate solutions are represented as organisms. Organisms are grouped into populations known as generations and are combined in pairs to produce subsequent generations. An individual organism's potential as a solution is determined by a fitness function. Fitness functions map organisms into real numbers and are used to determine which organisms will be used (and how frequently) to produce offspring for the succeeding generation. Fitness functions often require heuristic information because a precise measure of the suitability of a given organism (i.e., solution) is not always attainable. An example is the recognition (i.e., labeling) of segments in a scene. General characteristics of objects in the scene such as curvature, size, length, and relationship to each other may be known only within broad tolerance levels. That is, there is a great variability in the relationships among objects in different scenes.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jul 01, 1989
Accession Number
ADA230891

Entities

People

  • B. P. Buckles
  • C. A. Ankenbrandt
  • Frederick E. Petry
  • M. Lybanon

Organizations

  • United States Naval Research Laboratory

Tags

Communities of Interest

  • Materials and Manufacturing Processes

DTIC Thesaurus Topics

  • Abstracts
  • Algorithms
  • Classification
  • Computer Science
  • Exponential Functions
  • Fuzzy Logic
  • Fuzzy Sets
  • Gases
  • Genetic Algorithms
  • Gulf Stream
  • Mutations
  • Numbers
  • Oceans
  • Real Numbers
  • Recognition
  • Set Theory
  • Test And Evaluation

Readers

  • Approximation Theory.
  • Molecular Genetics
  • Neural Network Machine Learning.

Technology Areas

  • AI & ML
  • AI & ML - Machine Learning Algorithms
  • Biotechnology