Title :
Using evolutionary programming for modeling: an ocean acoustic example
Author_Institution :
Orincon Corp., San Diego, CA, USA
fDate :
10/1/1992 12:00:00 AM
Abstract :
The process of natural evolution is used as a basis for a search technique that can locate the extremum of complex response surfaces despite the existence of multiple local minima or maxima. Background on research in simulated evolution is offered. The fundamental properties of natural evolution are simulated for the purpose of modeling a set of ocean acoustic signals. The experimental results indicate that simulated evolution provides a method for estimating both the appropriate order and parameter values of a model. Some theoretical comparisons are made to standard estimation methods
Keywords :
acoustic signal processing; digital simulation; geophysics computing; oceanographic techniques; optimisation; parameter estimation; programming; search problems; underwater sound; ARMA models; complex response surfaces; evolutionary programming; extremum; modeling; multiple local maxima; multiple local minima; ocean acoustic signals; search technique; simulated evolution; underwater acoustics; Genetic programming; Linear approximation; Oceans; Parameter estimation; Predictive models; Response surface methodology; Sea measurements; Sea surface; System identification; Yield estimation;
Journal_Title :
Oceanic Engineering, IEEE Journal of