• Florent Pled, Université Gustave Eiffel, MSME UMR 8208
• Christophe Desceliers, Université Gustave Eiffel, MSME UMR 8208
• Maarten Arnst, Université de Liège, Aerospace and Mechanical Engineering

The numerical modeling, simulation and identification of random heterogeneous materials play an ever-growing role in material sciences and give rise to many appealing engineering and scientific challenges for the design of innovative metamaterials or the characterization of real-world existing materials, such as e.g. sedimentary rocks, natural composites, fiber- or nano-reinforced composites, some concretes and cementitious materials, some porous media, some living biological tissues, among many others. The robust identification of random material properties and models requires solving a statistical inverse problem that is usually formulated as a single- or multi-objective optimization problem and that can be solved using traditional gradient-based learning algorithms or global search algorithms, such as e.g. fixed-point iterative, random search, genetic and evolutionary algorithms, etc. This mini-symposium aims at presenting the recent developments on the statistical inverse identification and related stochastic optimization methods for the characterization of random heterogeneous materials. Data-driven scientific machine learning and probabilistic learning methods for identification purposes are also welcome. Practical applications and demonstration problems may concern, among others, linear and nonlinear random material models, involving e.g. plasticity, damage or fracture mechanisms or other material non-linearities, in uncertain linear and nonlinear computational mechanics for (quasi-)static or dynamic analyses.