Accelerated Newton-Raphson GRAPE methods for optimal control
David L. Goodwin, Mads Sloth Vinding
Abstract
Improvements to the optimal control Newton-Raphson GRAPE method for exact Hessian calculations are shown with examples of auxiliary matrix and ESCALADE methods. For an ensemble of two-level systems with realistic MRI conditions, the new accelerated Hessian calculations are between 200 and 600 times faster for auxiliary matrix and ESCALADE, respectively.
Topics & Concepts
Hessian matrixNewton's methodOptimal controlMatrix (chemical analysis)Quasi-Newton methodMathematicsApplied mathematicsHessian equationMathematical optimizationComputer sciencePhysicsMathematical analysisChemistryNonlinear systemChromatographyQuantum mechanicsFirst-order partial differential equationPartial differential equationAdvanced NMR Techniques and ApplicationsAdvanced MRI Techniques and ApplicationsNMR spectroscopy and applications