too many variables

[Michael A. Fetterman]

I'd like to be able to use gprolog to solve problems that look something like 
this:

goal(X1, X2, X3, X4, X5) :-
        X1 #<# X2,
        X2 #<# X3 * 2,
        X3 #<# X4 + 27,
        X5 #<# X1.

gprolog's FD variables are perfect for my application domain.

However, here's the catch.  I don't have 5 variables; I have 10000.

I've hit two problems so far.

First, I can't use goal(X1, X2, X3, ...) syntax because I hit the max_arity of
gprolog, which is currently set at 256.  I could change the #define and 
recompile,
but for my next attempt, I tried this instead:

goal([X1, X2, X3, X4, ...]) :-
        ...
        ...

This gets around the max_arity problem, but then I run into the max variables in
a term limit, which is currently set at 2048.

I could change that #define, and try again, but here's my question:

Am I missing the way in which I should be expressing this problem?  I need to
simultaneously solve a very large number of fairly simple relations like this.
Am I on the right track, or are these #define's in the gprolog source trying to
tell me something?

Thanks
Michael Fetterman