The model construction for phenomenological material response involves two main steps. First, it requires the formulation of a mathematical model to capture the physical effects. Then, the parameters of the underlying material model must be determined from experiments. The determination of material parameters therefore leads to an inverse problem, in which the unknown parameters must be fit to the experiments in an optimal sense. A classical approach to the identification of material parameters is the error minimization between model simulation and experimental data. This procedure leads to a highly-nonlinear optimization problem, in which the material parameters are the independent variables, known as parameter identification. The lecture offers an introduction to the basic concepts of experimental mechanics and parameter identification as well as nonlinear optimization with applications to selected model problems. Contents:

• Basic concepts of experimental material mechanics
• The inverse problem of parameter identification
• Nonlinear optimization methods and sensitivity analysis
• Gradient-based methods, evolution strategies, neural networks
• Finite element implementation of inhomogeneous problems
• Application to representative model problems