Customer Interviews:

Aytac Guven - Research in Hydraulic Structures

Aytac Guven works as a Research Assistant at the University of Gaziantep in Turkey. Mr. Guven's educational background extends from a Bachelors and Masters in Civil Engineering and is currently a doctoral candidate for Civil Engineering all from the University of Gaziantep.

Mr. Guven has been using NeuroSolutions for the past two years for the prediction and explicit formulation of local scour downstream of hydraulic structures (i.e. bridge piers,spillways, grade-control structures) as a function of inlow conditions, soil properties and scour geometry. Researchers have found that they are "saving a lot of time compared to other conventional numerical methods" and that "using neural network modeling is much more powerful than other prediction methods". Aytac generally achieves equal or greater than 99% correlation with experimental results!

The most commonly used neural network Mr. Guven uses is the Multi-Layer Percetron along with the second order learning algorithm Levenburg-Marquardt with one hidden-layer. In addition, Atyac almost always uses genetic optimization to find the optimum number of processing elements for his problems. Mr. Guven usually works with a single output variable with around 10,000 samples in his prediction problems.

Mr. Guven went on to say that "NS [NeuroSolutions] is the most user-friendly and comprehensive neural network tool I have ever used. Genetic algorithm features of NS [NeuroSolutions] empowers the efficiency of developed models". We would like to thank Mr. Guven for taking the time out of his busy schedule to discuss his success with NeuroSolutions and we hope to continue to hear great things in the future.