Flood. An open source neural networks C++ library

Abstract

The multilayer perceptron is an important model of neural network, and much of the literature in the �eld is referred to that model. The multilayer perceptron has found a wide range of applications, which include function re- gression, pattern recognition, time series prediction, optimal control, optimal shape design or inverse problems. All these problems can be formulated as variational problems. That neural network can learn either from databases or from mathematical models.

Flood is a comprehensive class library which implements the multilayer perceptron in the C++ programming language. It has been developed follow- ing the functional analysis and calculus of variations theories. In this regard, this software tool can be used for the whole range of applications mentioned above. Flood also provides a workaround for the solution of function opti- mization problems.