Record Details

A combined approach of genetic algorithm and neural networks with an application to geoacoustic inversion studies

NOPR - NISCAIR Online Periodicals Repository

View Archive Info
 
 
Field Value
 
Title A combined approach of genetic algorithm and neural networks with an application to geoacoustic inversion studies
 
Creator Yegireddi, Satyanarayana
 
Subject Underwater acoustic propagation
Inversion
ANN
Genetic algorithm
GANN
 
Description 195-201
Geoacoustic model of an offshore area derived
from inversion of pressure vector data using an appropriate forward model of
underwater acoustic propagation plays a significant role in sonar range
predictions and transmission loss estimation. Soft computing techniques like
Genetic Algorithms (GA), Artificial Neural Networks (ANN), Fuzzy logic have
definite advantages over the time consuming conventional approaches being used
in inversion, to solve complex non-linear phenomenon associated with
multi-dimensional parameter space. The present study deals with application of a
combined approach (GANN) of GA and ANN to geoacoustic inversion, exploiting the
merits of each technique. GANN found to be more promising and shows faster
convergence to a global minimum in optimisation of the inverted geoacoustic
parameters like density, compressional sound speed, attenuation and thickness
of the upper sediment strata of sea bottom within the tolerable error limits.
GANN shows better consistency in results unlike ANN. The approach is very
effective in faster simulation and takes a few minutes for inversion,
comparative to the traditional optimisation methods, which require a few hours.
 
Date 2016-07-04T08:35:54Z
2016-07-04T08:35:54Z
2015-02
 
Type Article
 
Identifier 0975-1033 (Online); 0379-5136 (Print)
http://hdl.handle.net/123456789/34633
 
Language en_US
 
Rights CC Attribution-Noncommercial-No Derivative Works 2.5 India
 
Publisher NISCAIR-CSIR, India
 
Source IJMS Vol.44(02) [February 2015]