Regularization of Nonlinear Least Squares Problems
We develop three of the five approaches described in Sect. 1.3.4 of the introduction: Levenberg–Marquardt–Tychonov (LMT), state-space, and adapted regularization. The two remaining approaches, regularization by parameterization and regularization by size reduction of the admissible parameter set, have been already addressed in Chaps. 3 and 4, respectively.
KeywordsRegularization Parameter Regularity Condition Tangent Cone Observation Operator Deflection Condition
- 1.Aki, K., Richards, P.G., 1980, Quantitative seismology: Theory and methods, W.H. Freeman, New YorkGoogle Scholar
- 8.Baumeister, J., 1987, Stable solutions of inverse problems, Vieweg, BraunschweigGoogle Scholar
- 18.Chavent, G., 1990, A new sufficient condition for the wellposedness of nonlinear least-squares problems arising in identification and control, In Bensoussan, A., and Lions, J.L., eds, Lecture Notes in Control and Information Sciences 144, Springer, Berlin, pp 452–463Google Scholar
- 26.Chavent, G., Kunisch, K., 1993, Regularization in state space, M2AN 27, 535–564Google Scholar