User-Defined Optimization: Matrix Estimation
Posted: Tue Sep 17, 2013 9:25 am
Hi,
I want to use the optimize command to estimate a matrix that simultaneously satisfies both row and column constraints. I know what the optimization functions look like for the rows and columns, and I know what I want to minimize, but I'm a little unsure of how to properly perform the solve.
I'm trying to figure out if I can code the row constraint (expressed as a function) into my called subroutine, and then specify what I want minimized, or alternatively, if I need to take the derivative of the optimization function, and then solve for what I want to be a minimum so I can reference this parameter in the called upon subroutine in the optimize command.
Secondly, I have a second optimization function to express the column constraints of the solved matrix. Is there a way to simultaneously satisfy both row and column constraints using user-defined optimization?
Many thanks,
Travis
I want to use the optimize command to estimate a matrix that simultaneously satisfies both row and column constraints. I know what the optimization functions look like for the rows and columns, and I know what I want to minimize, but I'm a little unsure of how to properly perform the solve.
I'm trying to figure out if I can code the row constraint (expressed as a function) into my called subroutine, and then specify what I want minimized, or alternatively, if I need to take the derivative of the optimization function, and then solve for what I want to be a minimum so I can reference this parameter in the called upon subroutine in the optimize command.
Secondly, I have a second optimization function to express the column constraints of the solved matrix. Is there a way to simultaneously satisfy both row and column constraints using user-defined optimization?
Many thanks,
Travis