[sci.math.symbolic] Vaxima out of core problem

rmadison@euler.Berkeley.EDU (Linc Madison) (07/17/89)

I've been having a rather frustrating time trying to get vaxima to help
me with a little differential equation I need to solve.  My equation is

	diff(x,t,2) + k*x +
	epsilon * ( c*diff(x,t) + f*signum(diff(x,t)) +
		    g*diff(x,t)*abs(diff(x,t)) + h*diff(x,t)^3
		  )                                              = 0

I need a solution to at least third order in epsilon, preserving the
identities of c,f,g,h in the solution.

I've tried pieces of this equation in ODE in vaxima, and also a sample
problem from the user's manual.  In all cases, I get a message about
improper arguments for ERROR and then quit with an "out of core" error
message.  My system parameters are: Ultrix-32 V3.0 (Rev 64); VAX 11/780
(?); Vaxima [MACSYMA] 2.11 in Franz Lisp opus 42, Wed Mar 19 12:44:59
1986; stacksize limit 512K, datasize limit 21984K, filesize, coredump-
size, and memoryuse unlimited.

I'm running out of people around Berkeley to ask, so if any wizardly
types could rescue me, I'd be most grateful.

The following is an edited transcript of a typical vaxima session with
the equations I'm working on.

-- Linc Madison   =   rmadison@euler.berkeley.edu

--------- VAXIMA TRANSCRIPT (...LONG...) FOLLOWS ------------

(d1) 				  vaxima1.out

(c2) load("ode/ode")$

/usr/local/mac/ode/ode.l being loaded.
[load /usr/local/mac/ode/ode.l]

(c3) load(ode2)$

/usr/local/mac/share/ode2.l being loaded.
[load /usr/local/mac/share/ode2.l]

(c4) depends(y,x);

(d4) 				    [y(x)]

/* This is a sample problem from the user manual */
(c5) eq:diff(y,x,2)*2*x*(1-x)+(1-x)*diff(y,x)+3*y;

				   2
				  d y	        dy
(d5) 		      2 (1 - x) x --- + (1 - x) -- + 3 y
				    2	        dx
				  dx

(c6) ode(eq,y,x);

Batching the file /usr/local/mac/ode/odeaux.v

(c7) /* -*- macsyma -*- */

[ ODE stuff edited out ]


					  log(x)	        log(x)
			  log(x)	  ------	        ------
			  ------	    2			  2
			    2	  3 log(%e       + 1)   3 log(%e       - 1)
(d19) y = 2 %k1 (x - 1) %e       (------------------- - -------------------
					   8			 8

									 log(x)
									 ------
				       3 x - 2				   2
			    - -------------------------) + %k2 (x - 1) %e
				  3 log(x)	 log(x)
				  --------	 ------
				     2		   2
			      4 %e	   - 4 %e

(c20) ode(eq,y,x,series);

Batching the file /usr/local/mac/ode/series.v

(c21) /* -*- Macsyma -*- */

/* series 71 */

[ more stuff edited out ]
(c29) ERROR(df,r1,r2,name,minsub):=block([],return(false))$
Warning - you are redefining the MACSYMA function error

[ more stuff edited out ]

(c69) variablep(u):=is(atom(u) and not numberp(u) or subvarp(u))$

(c70) zp1(n,zp):=block([z,p],z:first(zp),p:zp[2],
    if not constantp(n)
        then (if inpart(n,0) = "*"
		  then for j while inpart(n,j) # 'end do
			   (if not constantp(piece)
			        then (z:cons(piece = 0,z),
				      p:conssingularities(p,piece)))
		  else (z:cons(n = 0,z),p:conssingularities(p,n))),
    return([p,z]))$

Batching done.
gc:[*list:1270{68%}; fixnum:51{2%}; ut:69%]

/usr/local/mac/das/mstuff.o being loaded.
[fasl /usr/local/mac/das/mstuff.o]

error in subscript 
part fell off end.
  

Too few arguments supplied to error(df, r1, r2, name, minsub):

[out of core...sorry]

  

(c71) kill(dependencies);

(d71) 				     done

(c72) depends(x,t);

/* This is a piece of my equation */
(c74) eq:diff(x,t,2)+k*x+h*diff(x,t)^3;

			       2
			      d x      dx 3
(d74) 			      --- + h (--)  + k x
				2      dt
			      dt

(c75) ode(%,x,t);
gc:[*list:1280{68%}; fixnum:51{2%}; ut:63%]

Batching the file /usr/local/mac/ode/trans.v

[ more Vaxima stuff edited out ]

Batching done.

Batching the file /usr/local/mac/ode/lapl.v

[ more Vaxima stuff edited out ]

Batching done.

Batching the file /usr/local/mac/ode/diffac.v

[ more Vaxima stuff edited out ]

Batching done.

/usr/local/mac/ucb/cpoly.o being loaded.
[fasl /usr/local/mac/ucb/cpoly.o]

(%d, h, k)
polynomial not univariate   

error in subscript 
MACSYMA was unable to evaluate the predicate:

2 > []

  

Too few arguments supplied to error(df, r1, r2, name, minsub):

[out of core...sorry]

  

(c103) closefile();

---------- END OF VAXIMA TRANSCRIPT ----------

Thanks,
-- Lincoln Madison   =   rmadison@euler.berkeley.edu
		     =   ...!ucbvax!euler!rmadison
		     =   euler!rmadison@uunet.uu.net
If anyone wants to know what I hear, ask me; otherwise, I won't bother
to post a summary here.