| Author |
Comment/Response |
frustrated
|
 04/11/12 4:30pm
We are working on solving a partial differential equation with free boundary conditions using NDSolve. We enter the appropriate free boundary conditions (four of them), and Mathematica says the system is overdetermined, as there is only one dependent variable. We attempt to enter in only one boundary condition to see if this resolves the situation, but the error remains the same.
Our code is as follows:
parameters = {alpha -> 0.33, rho -> 0.1, sb -> 0.1, delta -> 0.1, ss -> 0.1, mu -> 0};
HJB1 = (alpha) k - rho V[k, s] + D[V[k, s], {k, 1}] (((s^2) - delta - mu))+0.5 ( s^2) (D[V[k, s], {k, 2}] + D[V[k, s], {s, 1}] (sb - s) + (0.5) D[V[k, s], {s, 2}] (ss^2)) == 0;
HJB2 = HJB1 /. parameters;
NDSolve[{HJB2, D[V[k, s], {k, 1}] == 0, D[V[k, s], {s, 1}] == 0, D[V[k, s], {k, 2}] == 0, D[V[k, s], {s, 2}] == 0}, V[k, s], {s, 0, 1}, {k, 0, 1}]
NDSolve::overdet: There are fewer dependent variables, {V[k,s]}, than equations, so the system is overdetermined. >>
Can anyone say what we're doing wrong?
URL: , |
|
