Greetings,
A few messages on this list from Jay Belanger in April 2002 state that:
[...]
> Desolve uses the LaPlace transform to solve odes, so it really only
> works well for linear odes.
[...]
> So desolve really only works with linear odes with constant
> coefficients.
I have a system of three linear DE's with constant coefficients, but
still get the result in terms of an inverse Laplace transform (ILT):
GCL (GNU Common Lisp) \ Version(2.4.1) Thu Apr \ 4 10:08:23 CST 2002
Maxima 5.6 Thu Apr 4 10:08:07 CST 2002 (with enhancements by W.
Schelter). [All on Linux/Intel]
EQN1:'DIFF(CA(T),T)-K10+CA(T)*K11-CB(T)*K12-CC(T)*K13=0;
EQN2:'DIFF(CB(T),T)-K20+CB(T)*K22-CA(T)*K21-CC(T)*K23=0;
EQN3:'DIFF(CC(T),T)-K30+CC(T)*K33-CA(T)*K31-CB(T)*K32=0;
ATVALUE(CA(T),T=0,CAi);
ATVALUE(CB(T),T=0,CBi);
ATVALUE(CC(T),T=0,CCi);
DESOLVE([EQN1,EQN2,EQN3],[CA(T),CB(T),CC(T)]);
[The result in terms of ILT and LVAR is very big so I won't give it here.]
My math is a bit rusty, but shouldn't this be pretty easy to solve? A
similar *two* equation system is solved with no problems. But it chokes
on the three coupled equations given above. If it's any consolation,
Mathematica also seems to choke above two equations (which may suggest I
am not specifying the problem correctly (?) ).
Neil
--
______________________________________________________
Neil E. Klepeis, UC Berkeley, School of Public Health,
and Lawrence Berkeley National Laboratory,
Berkeley, CA USA. Voice: 831-768-9510