Parallel genetic evolution of membership functions and rules for a fuzzy controllerCastellano GHIGH-PERFORMANCE COMPUTING AND NETWORKINGLECTURE NOTES IN COMPUTER SCIENCE

bankpottstownΤεχνίτη Νοημοσύνη και Ρομποτική

23 Οκτ 2013 (πριν από 3 χρόνια και 11 μήνες)

62 εμφανίσεις

Parallel genetic evolution of membership functions and rules for a fuzzy
controller

Mondelli G,
Castellano G
, Attolico G, Distante C

HIGH
-
PERFORMANCE COMPUTING AND NETWORKING


LECTURE NOTES IN COMPUTER SCIENCE


1401: 922
-
924 1998


Document
type:

Article



Language:

English



Cited
References:

4



Times
Cited:

0




A
bstract:

Parallel Genetic Algorithms (PGAs), implemented on the APE100/Quadrics SIMD
architecture, were applied to automatic design of membership functions and fuzzy
rules for robotic control. They run multiple simultaneous searches, differently
balancing
exploration of the solution space and fine tune of the best solutions
available at each generation. Migration spreads the best individuals of each population
in local neighborhoods. The approach reduces the time required for fitness evaluation
(each popula
tion has less individuals), decreases the generations required for
acceptable solutions and increases the probability of identifying optimal solutions.

Addresses:

Mondelli G,
CNR
, Ist Elaboraz Segnali & Immagini, Via Amendola 166
-
5, I
-
70126
Bari, Italy

CN
R
, Ist Elaboraz Segnali & Immagini, I
-
70126 Bari, Italy

Publisher:

SPRINGER
-
VERLAG BERLIN, BERLIN

IDS Number:

BM08P

ISSN:

0302
-
9743