how to ignore "Unable to compute the LU factorization"

Two ideas:
(1) instert an errcatch into the mnewton code.
(2) change the mnewton code to use rational numbers for the LU
factorization.

Barton

-----maxima-bounces at math.utexas.edu wrote: -----

>To:?maxima at math.utexas.edu
>From:?Lester?Ingber?<ingber at ingber.com>
>Sent?by:?maxima-bounces at math.utexas.edu
>Date:?04/04/2009?11:02AM
>Subject:?[Maxima]?how?to?ignore?"Unable?to?compute?the?LU?factorization"
>
>I'm?running?loops?over?a?two-variable?nonlinear?set?of?two?equations
>using?mnewton,?e.g.,
>ELresults_rl:mnewton([eq_e_rl,eq_i_rl],[w_e,w_i],[e_ini,i_ini])
>which?is?embedded?in?a?loop?over?other?variable?values.??I.e.,?it?is
>being?called?many?times.
>
>This?works?pretty?well?most?of?the?time.??However,?once?in?awhile?I?get
>Unable?to?compute?the?LU?factorization
>#0:?solve_by_lu(
>Maxima?encountered?a?Lisp?error:
>and?then?I?get?kicked?out?of?the?loop?over?mnewton,?not?even?trying?all
>the?other?values.??All?hell?breaks?loose,?and?depending?on?the?machine
>and?lisp?version?I'm?using?I?get?a?string?of?other?lisp?complaints.
>(I.e.,?I?seem?to?get?the?same?problems?at?the?same?values,?so?yes?this
>is?some?problem?being?fed?to?mnewton?by?the?algebra?in?the?equations
>developed?by?my?problem.)
>
>I'd?like?to?be?able?to?simple?ignore?these?cases?produced?by?some?merror
>in?an?LU?function.??E.g.,?the?results?are?fed?to?a?file?that?is?processed
>by?gnuplot,?so?NaN's?are?just?fine.??Is?there?any?setting?I?can?give?to
>my?batch?jobs?to?do?this?
>
>Thanks.
>
>Lester
>
>
>_______________________________________________
>Maxima?mailing?list
>Maxima at math.utexas.edu
>http://www.math.utexas.edu/mailman/listinfo/maxima