Lying by Approximation

In teaching an introduction to the finite element method at the undergraduate level, a prudent mix of theory and applications is often sought. In many cases, analysts use the finite element method to perform parametric studies on potential designs to size parts, weed out less desirable design scenarios, and predict system behavior under load. In this book, we discuss common pitfalls encountered by many finite element analysts, in particular, students encountering the method for the first time. We present a variety of simple problems in axial, bending, torsion, and shear loading that combine the students' knowledge of theoretical mechanics, numerical methods, and approximations particular to the finite element method itself.

This book discusses soft computing, which provides an efficient platform to deal with imprecision, uncertainty, vagueness and approximation in order to attain robustness and reliable computing. It explores two major concepts of soft computing: fuzzy set theory and neural networks, which relate to uncertainty handling and machine learning techniques respectively. Generally, fuzzy sets are considered as vague or uncertain sets having membership function lying between 0 and 1, and ANN is a type of artificial intelligence that attempts to imitate the way a human brain works by configuring specific applications, for instance pattern recognition or data classification, through learning processes.