tikhonov regularization solution looks very similiar to NNLS solution.

Hello,

I am using Tikhonov regularization with non negativity constraints that i am implementing with the lsqnonneg function. I am also using the L curve function provided by regtools to determine the optimal degree of regularization. However, both of these methods provided a solution that is very similar to the solution obtained with ordinary NNLS without regularization. I am wondering if anyone has encountered a similar problem and has suggestions on possible solutions. I am thinking about using the second derivative operator in lieu of the identity matrix in the Tikhonov as well as implementing a truncated SVD. Thanks,

Mike.