We are trying to solve NLE using NLPTR subroutine in IML. We have set the following control parameters:
par = j(1,10,.); par=8; par=.00001; par = 0.0000001;
We are getting a very large step size in the very first iteration that ends up hitting upper bound constraint. The optimization quits as it hits the upper constaint. Is there any other way of controlling the step size? I quickly wrote the same code in matlab and verified that the solution does in fact falls between the two constraints that I have specified.