Seminar WS 22/23 Graduate Seminar on Efficient Simulation
Hierarchical Matrices
Many problems in science and engineering require to perform numerical computations with dense matrices which lead to high computational cost and memory requirements which are prohibitive for larger problem sizes. However, many dense matrices provide an underlying analytical structure which allows for a data-sparse approximation by hierarchical matrices and H2-matrices in linear or almost linear cost. In the seminar we will discuss the basic principles of hierarchical and H2-matrices as well as efficient algorithms for algebraic operations.