PIMS-SFU CSC Seminar -Roland Herzog

Optimization problems with sparsity enhancing norms have helped enable the tremendous success of computational compressed sensing techniques in the past decade, where the goal is to recover a signal from a small number of measurements. In this presentation we consider sparsity enhancing norms from a different perspective. In the context of optimal control problems they can be used to promote sparse controls, which has immediate applications in actuator placement problems. The talk introduces problem settings with L1 and measure controls as well as directional (joint) sparsity variants. We touch upon analytical aspects, efficient algorithms, numerical analysis and examples.