Number of variables, equations and arrais
Posted: Mon Apr 13, 2009 1:56 am
Dear Dynare experts,
I have model with many sectors and many regions.
I'm migrating from GAMS. The model in GAMS is very compact because GAMS has pre-processor which generates final equations based on 'sets'. As I know, Dynare/Matlab doesn't allow to have sets or arrays. And if I have similar set of equations for each region and sector, there is no other ways to code it then write it personally for each region. Please correct me if I'm wrong. My code is too large to be written manually. And I'm writing code in GAMS that generates code for DYNARE, may be you can suggest me a way to compact the code.
Another question:
How many endogenous variables and equations may I have? I found here in the forum similar question with answer that number of endogenous variables may be up to few hundred. But the information goes from 2007. Are there any changes by now?
Please share your experience, what is the biggest model you have run with DYNARE, and how much time it takes to estimate/solve the model?
Thanks,
Oleg
I have model with many sectors and many regions.
I'm migrating from GAMS. The model in GAMS is very compact because GAMS has pre-processor which generates final equations based on 'sets'. As I know, Dynare/Matlab doesn't allow to have sets or arrays. And if I have similar set of equations for each region and sector, there is no other ways to code it then write it personally for each region. Please correct me if I'm wrong. My code is too large to be written manually. And I'm writing code in GAMS that generates code for DYNARE, may be you can suggest me a way to compact the code.
Another question:
How many endogenous variables and equations may I have? I found here in the forum similar question with answer that number of endogenous variables may be up to few hundred. But the information goes from 2007. Are there any changes by now?
Please share your experience, what is the biggest model you have run with DYNARE, and how much time it takes to estimate/solve the model?
Thanks,
Oleg