I'm estimating a model using simulated methods of moments and I'm using optmum to find the minimum distance between the true and the simulated moments. I'm telling Gauss to print the vector of parameters at the beginning of every iteration (or trials within iterations). I've noticed that sometimes Gauss returns a vector of missing values as the update on the parameter vector. Does anyone know what that means? I've checked if the Hessian is invertible and whether the direction vector exist and I cannot find a problem there. I cannot think of a reason why the update would be an empty vector. Any help would be highly appreciated!!
During the iterations there are times when one of a number of catastrophic failures occur. Optmum is designed to recover from them if possible. If it can't recover, Optmum will return a return code describing the failure. I would need to be able to run your example in order to say what exactly failed.