Development Of Wave Buoy Network Using Soft Computing Techniques

Created by W.Langdon from gp-bibliography.bib Revision:1.7630

@InProceedings{Londhe:2008:OCEANS,

• author = "Shreenivas N. Londhe",
• title = "Development Of Wave Buoy Network Using Soft Computing Techniques",
• booktitle = "OCEANS 2008 - MTS/IEEE Kobe Techno-Ocean",
• year = "2008",
• month = "8-11 " # apr,
• pages = "1--8",
• address = "Kobe, Japan",
• keywords = "genetic algorithms, genetic programming, AD 2002 to 2004, Artificial Neural Networks, Australia, Canada, Germany, Gulf of Mexico, India, UK, USA, buoy programs, ocean wave buoys network development, ocean wave data measurements, soft computing techniques, stochastic techniques, geophysics computing, neural nets, ocean waves, oceanographic techniques, stochastic processes",
• DOI = "doi:10.1109/OCEANSKOBE.2008.4530913",
• abstract = "Wave buoys are perhaps the only reliable source measuring waves continuously for years. This is perhaps the most vital reason for establishment of data buoy programs by various countries like USA (NDBC), Australia, Canada, UK, Germany, India (NDBP) etc. The wave data measurements not only provide real time wave information for Coastal and Ocean related activities but also form wave data base useful for predicting future events using statistical or stochastic techniques. However some times these wave buoys stop functioning either due to malfunctioning instruments or maintenance-related reasons resulting into loss of data. This paper presents use of soft computing techniques like Artificial Neural Networks (ANN) and Genetic Programming (GP) to retrieve this lost data by forming a network of wave buoys in a region. For developing the buoy network common data of hourly significant wave heights at six buoys in the Gulf of Mexico namely 42001, 42003, 42007, 42036, 42039 and 42040 for the years 2002 and 2004 is used. A separate network for each buoy is developed as the 'target buoy' with other 5 buoys as 'input buoys' which can be operated to retrieve lost data at a location. The testing results of both approaches when compared showed superiority of Genetic Programming over Artificial Neural Network as evident by higher correlation coefficient between observed and predicted wave heights in all cases. The wave height plots also pointed out that GP estimates wave heights in extreme events (peaks) more accurately than ANN.",
• notes = "Also known as \cite{4530913}",

}

Genetic Programming entries for S N Londhe

Citations