- #1
totally_lost
- 2
- 0
Dear all,
I have a PDF in independent variables q, [tex]\mu[/tex] and h, all depending on x and y. I wish to check whether or not this PDF converges, which means checking that the normalisation constant converges in the limits -[tex]\infty[/tex] and +[tex]\infty[/tex] of the above mentioned variables q, [tex]\mu[/tex] and h. The integral is given by
C[tex]^{-1}[/tex] = [tex]\int[/tex] dx dy [tex]\int[/tex][tex]\int[/tex][tex]\int[/tex] dq dh d[tex]\mu[/tex] h3 e-1/2[tex]\beta[/tex](g h2+h(k x [tex]\nabla[/tex]InvLap(h q - f) + [tex]\nabla[/tex]InvLap[tex]\mu[/tex])2) - h [tex]\sum[/tex] [tex]\alpha[/tex]j qj .
Note that q = q(x,y), h = h(x,y) and [tex]\mu[/tex] = [tex]\mu[/tex](x,y), k = (0, 0, 1) and InvLap is the inverse laplacian operator. Beta and alpha_j are lagrange multipliers which may still be scaled freely. The sum runs from j=0 to j= K < infinity.
Any ideas on how to tackle this problem in terms of proving convergence are welcome.
I have a PDF in independent variables q, [tex]\mu[/tex] and h, all depending on x and y. I wish to check whether or not this PDF converges, which means checking that the normalisation constant converges in the limits -[tex]\infty[/tex] and +[tex]\infty[/tex] of the above mentioned variables q, [tex]\mu[/tex] and h. The integral is given by
C[tex]^{-1}[/tex] = [tex]\int[/tex] dx dy [tex]\int[/tex][tex]\int[/tex][tex]\int[/tex] dq dh d[tex]\mu[/tex] h3 e-1/2[tex]\beta[/tex](g h2+h(k x [tex]\nabla[/tex]InvLap(h q - f) + [tex]\nabla[/tex]InvLap[tex]\mu[/tex])2) - h [tex]\sum[/tex] [tex]\alpha[/tex]j qj .
Note that q = q(x,y), h = h(x,y) and [tex]\mu[/tex] = [tex]\mu[/tex](x,y), k = (0, 0, 1) and InvLap is the inverse laplacian operator. Beta and alpha_j are lagrange multipliers which may still be scaled freely. The sum runs from j=0 to j= K < infinity.
Any ideas on how to tackle this problem in terms of proving convergence are welcome.