Appendix A: Conjugate gradient minimization of the Rayleigh Quotient

Recall that $\bold q = [\bold m \;\; -1]^T$, where m is the ``usual'' model (i.e., Lm=d). $\bold q_i$ is the estimated model vector at iteration i.

$$\begin{eqnarray} \bold q_0 &=& \frac{\bold q_0}{\sqrt{\bold q_0^T \bold q_0}} \h... ...ter}} &=& \bold q_{n_{iter}}/\left(-\bold q_{n_{iter}}[m+1]\right)\end{eqnarray}$$ (131) (132) (133) (134) (135) (136) (137) (138) (139) (140) (141) (142) (143) (144) (145) (146) (147) (148) (149) (150) (151)

 

 

Short Note
Nonlinear pairwise alignment of seismic traces

Christopher L. Liner and Robert G. Clapp

cll@utulsa.edu, bob@sep.stanford.edu