On Variant Strategies to Solve the Magnitude Least Squares Optimization Problem in Parallel Transmission Pulse Design and Under Strict SAR and Power Constraints

摘要: Parallel transmission is a very promising candidate technology to mitigate the inevitable radio-frequency (RF) field inhomogeneity in magnetic resonance imaging at ultra-high field. For first few years, pulse design utilizing this technique was expressed as least squares problem with crude power regularizations aimed controlling specific absorption rate (SAR), hence patient safety. This approach being suboptimal for many applications sensitive mostly magnitude of spin excitation, and not its phase, (MLS) then formulated 2007. Despite importance availability other powerful numerical optimization methods, MLS yet has been faced almost exclusively by designer so-called variable exchange method. In paper, we investigate various two-stage strategies consisting different initializations nonlinear programming approaches, incorporate directly strict SAR hardware constraints. Several schemes such sequential quadratic programming, interior point semidefinite squared relaxations are studied both small large tip angle regimes RF static maps obtained vivo on human brain 7T. Convergence robustness approaches analyzed, recommendations tackle finally given. Small inversion pulses returned seconds under minute respectively while respecting constraints, allowing use proposed routine.