Calibrating artificial neural networks by global optimization

Abstract

An artificial neural network (ANN) is a computational model − implemented as a computer program − that is aimed at emulating the key features and operations of biological neural networks. ANNs are extensively used to model unknown or unspecified functional relationships between the input and output of a "black box" system. In order to apply such a generic procedure to actual decision problems, a key requirement is ANN training to minimize the discrepancy between modeled and measured system output. In this work, we consider ANN training as a (potentially) multi-modal optimization problem. To address this issue, we introduce a global optimization (GO) framework and corresponding GO software. The practical viability of the GO based approach is illustrated by finding close numerical approximations of (one-dimensional, but non-trivial) functions.

The Performance Analysis of Artificial Neural Network Based Shimizu-Morioka Chaotic System with Respect to Sample Numbers

“…It is possible to use ANN in function estimation, classification, pattern recognition, signal processing and system modeling. ANN have recently been used in many fields like prediction [1,2], classification [3,4], control [5,6], optimization [7,8], the analysis of complex problems and modeling of nonlinear systems [9,10]. The usage of ANN in modeling of nonlinear systems has an important role in modeling chaotic oscillators using ANN.…”