Next: About this document ... Up: Brown: Total least-squares Previous: REFERENCES

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.

$\begin{eqnarray} \bold q_0 &=& \frac{\bold q_0}{\sqrt{\bold q_0^T \bold q_0}} \h... ...ter}} &=& \bold q_{n_{iter}}/\left(-\bold q_{n_{iter}}[m+1]\right)\end{eqnarray}$ (7)

(8)

(9)

(10)

(11)

(12)

(13)

(14)

(15)

(16)

(17)

(18)

(19)

(20)

(21)

(22)

(23)

(24)

(25)

(26)

(27)

Next: About this document ... Up: Brown: Total least-squares Previous: REFERENCES

Stanford Exploration Project
11/11/2002

	$\begin{eqnarray} \bold q_0 &=& \frac{\bold q_0}{\sqrt{\bold q_0^T \bold q_0}} \h... ...ter}} &=& \bold q_{n_{iter}}/\left(-\bold q_{n_{iter}}[m+1]\right)\end{eqnarray}$	(7)
		(8)
		(9)
		(10)
		(11)
		(12)
		(13)
		(14)
		(15)
		(16)
		(17)
		(18)
		(19)
		(20)
		(21)
		(22)
		(23)
		(24)
		(25)
		(26)
		(27)