报告题目：Artificial intelligence in computational mechanics: towards manifold learning approach

报 告 人：Pro. Piotr Breitkopf

时    间：2018年3月27日（周二）下午14:00

地    点：学科2号楼报告厅

报告内容：

In this contribution, we present the concept of a "shape manifold" designed for reduced order representation of complex 'shapes' encountered in mechanical problems, such as design optimization, springback or image correlation. The overall idea is to define the manifold of admissible shapes within which evolves the boundary of the structure. The reduced representation is obtained by means of projecting the level set representation of shapes on a set of carefully chosen basis vectors. This allows us to identify the intrinsic dimensionality of the problem, independently of the original design parameters. Also, an optimal parameterization may be obtained for arbitrary shapes, where the parameters have to be defined. We also developed the predictor-corrector optimization algorithms in a reduced shape space that guarantee the admissibility of the solution with no additional constraints. We illustrate the approach on three diverse examples drawn from the field of computational and applied mechanics.

专家简介：

    Piotr Breitkopf教授现任法国国家实验室常务副主任、Roberval实验室主任，是法国国家科学院第九区委员、计算结构力学联盟（CSMA）执行委员、Computer Science编委，主要研究成果包括大规模优化设计问题建模与缩减模型优化算法等。承担欧盟和法国政府及企业项目10余项，发表学术论文200余篇，编写学术专著9部。