Abstract

The UMN proposal aims at re-enforcing the existing network initiated by the «Unfold Mechanics for Sounds and Music» colloquium organized at IRCAM in September 2014. Since then, some experts have agreed to enhance the quality of the consortium (see members below) and an international committee have also been contacted by the coordination in order to facilitate the access to the European funding programmes (Horizon 2020, FET). With the High Performance Computing (HPC) challenge in mind, this European scientific network will use the MRSEI ANR instrument to organize research addressing geometric methods (in a broad sense) in mechanics and control theory.

Geometric methods for designing structurally sound algorithms and simulation will be investigated along two key directions: symmetry and modularity. The first direction yields the celebrated Noether’s theorem which relates the symmetries to the existence of conserved quantities. This often allows for considerable reductions and divides the computational complexity by several orders of magnitude.

On the other hand, dealing with “multibody systems” involving interface connections, the port Hamiltonian approach is particularly well suited. It goes without saying that modularity and its associated symmetries may drastically reduce the computational times, and also, minimize data movement in extreme computing and boost parallelism of simulation codes.

Based on a pre-existing team of multidisciplinary eminent experts, the UMN network is now consolidated by an HPC platform (ICS) and an industrial partner. The MRSEI instrument will help to make mathematicians, engineers, computer scientists and leaders in the industry cooperate to ensure European leadership in the supply and use of HPC systems and services by 2020.

4 thoughts on “Abstract

  1. Tudor Ratiu

    Since Darryl already mentioned port Hamiltonian systems the obvious next question is how to incorporate in this stochastic and numerical methods. I know that Hong Wang in Tianjin has worked on the geometric side of control theory, linked to Hamiltonian systems, but I am not sure how far she got and how deep her results really are. I will try to look into this more carefully.

    Reply

Leave a Reply to Holm Cancel reply

Your email address will not be published. Required fields are marked *