Abstract. The PUMA software toolkit is designed to allow for the rapid development of simulation applications using generalized finite element techniques based on the partition of unity method (PUM). Unlike classical finite element methods (FEM) a PUM can directly utilize user insight, domain-specific information and physics-based basis functions to reduce the computational cost substantially and thereby allows for the rapid evaluation of novel models. We discuss the basic building blocks of the PUMA software framework and present some examples showcasing the capabilities of PUMA and its ease of use. To this end, we consider laminated composite plates and discuss how small scale structural details that are typically ignored in design and margin of safety applications due to excessive computational costs when using classical approximation methods such as finite elements. Here, we present the construction of optimal basis functions that implicitly resolve these details so that it becomes feasible to consider these details also in design.

Bio: Marc Alexander Schweitzer is the managing director of the Institute for Numerical Simulation at the University of Bonn and holds the chair for Scientific Computing. Moreover, he is the head of the department Numerical Software at the Fraunhofer Institute for Algorithms and Scientific Computing SCAI where he is responsible for e.g. the software products SAMG, MESHFREE, PUMA, and OpenCoSim. Previously, he was at the University of Stuttgart and held the chair for Simulation of Large Systems at the Institute for Parallel and Distributed Systems. He received a Diplom in Mathematics, a Doctorate in the Sciences and Habilitation from the University of Bonn.