Neuroevolutionary optimization

TL;DR

This paper explores how evolutionary algorithms can optimize artificial neural networks by evolving weights, architectures, and learning rules, analyzing key issues and proposing effective strategies for parameter optimization.

Contribution

It provides a comprehensive review of different evolutionary approaches for neural network optimization and discusses critical challenges and solutions.

Findings

Evolution of connection weights, architectures, and learning rules are distinct but interconnected processes.

Effective strategies for using evolutionary algorithms in neural network parameter optimization are identified.

Critical issues in applying evolutionary methods to neural networks are analyzed and addressed.

Abstract

This paper presents an application of evolutionary search procedures to artificial neural networks. Here, we can distinguish among three kinds of evolution in artificial neural networks, i.e. the evolution of connection weights, of architectures, and of learning rules. We review each kind of evolution in detail and analyse critical issues related to different evolutions. This article concentrates on finding the suitable way of using evolutionary algorithms for optimizing the artificial neural network parameters.