You can use dsolve to solve a system of differential equations:

eq1:=(diff(g(t), t, t))*f(x)-(diff(f(x), x, x))*g(t) = 0:

eq2:=(diff(f(x), x))^2*g(t)-(diff(g(t), t))^2*f(x) = 0:

sys:=[eq1,eq2]:

dsolve(sys);

the numerical pde solver can't handle 3 variables, as far as I know, so unless you can transform the system somehow, there may be little you can do. The equation you posted above looks a little odd - what are H0, lambda0 and Q ?

What makes you think that Maple can't do it ?

A lot will depend on your boundary and initial conditions, but in general, I think psdolve should generate a general solution which incorporates some constants and ODEs which you can then solve with dsolve.

Edit: the above applies to a linear pde. It seems the OP's problem is nonlinear, so there may be no analytic solution.

Some of the more experienced users may be able to provide better help, but you might also find this useful

http://www.mcs.anl.gov/~wozniak/papers/wozniak_mmath.pdf

You could combine those steps into one:

tf3:=TransferFunction(tf1:-tf[1,1]*tf2:-tf[1,1]);

Does that help ?

Hi, maybe I misunderstood the point of your question, but does subs do what you want ?
subs(L[a]=0,deq);

? 

restart:

d:=1/( sqrt(a^2+p^2) ):

aa:= sin(theta + tao*s)*cos(d*s) + p*d*cos(theta + tao*s)*sin(d*s):

bb:=  sin(theta + tao*s)*sin(d*s) - p*d*cos(theta + tao*s)*cos(d*s):

cc:= a*d*cos(theta + tao*s):

hr:=sqrt ( aa^2 +bb^2 +cc^2 );

simplify(%) assuming theta::real,tao::real,s::real,a::real,p::real;

Maybe this will help

http://www.mapleprimes.com/posts/36308-Classic-On-XP64

Please check the equations, they are not in proper Maple notation. Are these correct ?
aa:= sin(theta + tao*s)*cos(d*s) + p*d*cos(theta + tao*s)*sin(d*s); 
bb:=  sin(theta + tao*s)*sin(d*s) - p*d*cos(theta + tao*s)*cos(d*s);
cc:= a*d*cos(theta + tao*s); 

?

Well, this isn't pretty, but I think it works, at least for the examples used in this thread so far. I guess someone will find an example it doesn't work for :)

rdiscont:=proc(f,x)
local tmp,rd,i,b:
rd:=NULL:tmp:=discont(f,x):
for i from 1 to nops(tmp) do  
if type(tmp[i], realcons) then rd:=rd,tmp[i] 
 elif tmp[i]=Re(tmp[i]) then rd:=rd,tmp[i]
 elif nops(indets(tmp[i]))>0 then  
   b:=eval(tmp[i],indets(tmp[i])[1]=a);  
   if not type(limit(b,a=0),undefined) then    
     if type(eval(b,a=0),realcons) then rd:=rd,eval(b,a=0) end if;
   end if;
 end if;
end do;
{rd};
end proc:

f := 3/(1-exp(1/x)):
rdiscont(f,x);
                                     {0}

f:=x/sin(x):
rdiscont(f,x);
                                  {Pi _Z21}

f:=1/(exp(x)-2):
rdiscont(f,x);
                                   {ln(2)}

perhaps fdiscont would work ?

If you want a perfect fit then, and you are fitting a polynomial to the data, then if you have n data points, you need a polynomial of degree n-1

In statistics this is referred to as overfitting and is usually considered bad practice, since you are generally trying to model the data, and a perfect-fitting model is not really a model at all and of very limited use for inference.

First you need to replace e() with exp()

Then, I think you have to  re=formulate the problem. I may be wrong but I don't think you can have coupled boundary conditions. If you want a unique solution I think you need 4 independent conditions. Can you not translate this system so that the coupled boundary conditions give a new equation with one fixed boundary/initial condition, then you have will have a system of three equations in two unknowns, with four conditions - which may then be possible to solve.

I had to submit some coursework yesterday and thought I'd try exporting to PDF from my friend's Maple 14. I'm sorry to say, that the results were not close to being acceptable ! So I did my usual "trick" of exporting to rtf and editing it in Word.

Probably the best thing to do is contact support

www.maplesoft.com/support  or depending on where you bought it, your local provider.