Numerical Instability?

GAMS
1

Step 1:
MODEL A: I solved a cost minimization problem using NLP. The results
for the endogenous variables Y look okay.

Step 2:
Then, I included one more equation (W) which calculates a new variable
(X).
The variable X is an expression based on the endogenous variables (Y)
of Model A.

This new equation (W) is included with Model A and solved – MODEL B.
In the presence of the equation W, the solution for endogenous
variables (Y) is
drastically different and does not make sense.

Why don’t the results for Model B make sense?
Is it due to numerical instability - how to correct it?

Thank you.
Subbu

\

2

Please note that equation W is NOT a Constraint affecting the
calculation of endogenous variables Y.

\

3

Subbu,

You are quiet right. This behavior as you describe does not make
sense. But there is probably a good explanation if one looks at the
details. I think you would need to share the model with us to get more
help.

Regards,
Michael Bussieck - GAMSWorld Coordinator

On Jun 11, 2:54 pm, whoami wrote:

Please note that equation W is NOT a Constraint affecting the
calculation of endogenous variables Y.

\