ABSTRACT
This chapter discusses the fundamental differences between hard computing and soft computing. The soft computing methods that considers in the chapter are based on adaptive learning systems, specifically neural networks. The chapter also explores the application of neural networks in material modeling. Although neural networks and genetic algorithms have been used in the modeling, design, and condition monitoring of structural systems, the application of neural networks in material modeling is potentially an important development in computational mechanics. Hard computing methods are generally mathematically based; they inherit their characteristic properties from the mathematics. These properties include precision, universality, and functional uniqueness. Soft computing methods such as neural networks, genetic algorithms, fuzzy logic, and swarm intelligence are biologically inspired. The chapter describes a method for modeling hysteretic behavior of materials. The autoprogressive algorithm is a method for training a neural network to learn the constitutive properties of materials from the results of structural tests.
