Du är här

Evolving Programs and Solutions Using Genetic Programming with Application to Learning and Adaptive Control

Författare:
Publiceringsår: 2002
Språk: Engelska
Sidor: 289-307
Publikation/Tidskrift/Serie: Journal of Intelligent and Robotic Systems
Volym: 35
Nummer: 3
Dokumenttyp: Artikel
Förlag: Kluwer

Sammanfattning

This paper discusses two feasibility studies of Genetic Programming (GP) to the field of control theory, GP being a method inspired from nature where the goal is to create a computer program automatically from high-level statements of problems' requirements. The first feasibility study derives from stability theory and deals with evolving a program that can solve discrete-time Lyapunov equations. The second application of GP tackles the problem of producing a self-evolved Model Reference Adaptive System (MRAS). Basic structure of the programs used in the experiments are only marginally different, yet applied to seemingly quite different problems. In the first feasibility study, it was observed that GP, beside correct usage of global variables, could also purposely arrange mathematical functions and operations in an iterative manner without being explicitly programmed for the task. In the second feasibility study, a controller was evolved for a second-order process based on a pre-defined reference model.

Disputation

Nyckelord

  • Technology and Engineering
  • model reference adaptive systems
  • learning systems
  • adaptive control
  • genetic programming
  • Lyapunov functions

Övriga

Published
Yes
  • ISSN: 0921-0296

Box 117, 221 00 LUND
Telefon 046-222 00 00 (växel)
Telefax 046-222 47 20
lu [at] lu [dot] se

 

Fakturaadress: Box 188, 221 00 LUND
Organisationsnummer: 202100-3211
Om webbplatsen