Numerical Analysis Seminar: Time adaptive waveform iterations for coupled problems
Niklas Kotarsky, Title: Time adaptive waveform iterations for coupled problems
Coupled problems like for example a feather moving in the wind or hot steel that is quenched in water are often difficult to simulate in an energy efficient and practical manner. The goal is to develop a method for simulating coupled problems that is high order and adaptive in time. Furthermore, the method must allow different time steps in the separate sub problems, as well as the reuse of old codes for the sub problems. In addition, all involved iterations should converge fast. This can be achieved by using so called waveform iterations, resulting in a fix-point iteration. The Quasi Newton method is often used to accelerate the convergence of the fix-point iteration. It has also recently been combined with waveform iterations to extend the Quasi Newton method to work with multirate time stepping. In this talks I will discuss some of the convergence properties of the waveform iterations as well as how to extend the Quasi Newton method to the time adaptive setting.
Tid: 2022-12-12 15:00 till 16:00
alexandros [dot] sopasakis [at] math [dot] lth [dot] se