DNLP or NLP and plausible Declaration??

[Archiver]

Hallo,

i have questionable Situation(:-)).

i have a Function Y(x^2, P1,P2, Q, Price)
with:
-P1(c,o), P2(c,o) and Q(c,o) are Parameters
-X(c,o) is Variable 1
-Price(c) is Variable 2.
With Constrains 1,2,3 i would like to delimit X:
1) X(c,o)=1 if Q(c,o)*(P1(c,o)-Price(c)) >=0
2) X(c,o)=0 if Q(c,o)*(P2(c,o)-Price(c)) 0))..
x(c,o)*(P1(c,o) -P2(c,o))=e= (Price(c)-P2(c,o));

and i get errror message
-with CPLEX (qcp): Endogenous relational operations require model type
dnlp
-WIth DNLP: endogenous $-control operations is not allowed

Can someone suggest why this is and what i can make?

many thanks in advance!
Dieter


--~--~---------~--~----~------------~-------~--~----~
To post to this group, send email to gamsworld@googlegroups.com
To unsubscribe from this group, send email to gamsworld+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/gamsworld?hl=en
-~----------~----~----~----~------~----~------~--~---

[Archiver]

Hello,
many Thanks an Ervin.

can someone suggest how i can make for CPLEX.

i write in GAMS
*--------------

file fcplex Option file /cplex.opt/;
putclose fcplex 'indic Constr_1(c,o)$((Q(c,o)*(P(t,c,o)-Price(c))) ge
0)'/
'indic Price_2(c,o)$((Q_(c,o)*(P_u(t,c,o)-Price(c)) le
0)';
mod1.optfile=1;

loop((t,c,o),
put fcplex 'indic' Constr_1.tn(c,o)'$'x.tn(c,o) '$'P(c,o)'$'Q
(c,o)'$Price(c) yes/
'indic' price_2.tn(t,c,o)'$$'x.tn(c,o) '$'P
(c,o)'$'Q(c,o)'$Price(c) yes/
);
putclose fcplex
Option miqcp=cplex;

$offtext
solve mod1 using dnlp maximizing WElfare_global;

*---------------------------------

and i agree the X over all c and o =1. this is false...



thanks
Dieter
--~--~---------~--~----~------------~-------~--~----~
To post to this group, send email to gamsworld@googlegroups.com
To unsubscribe from this group, send email to gamsworld+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/gamsworld?hl=en
-~----------~----~----~----~------~----~------~--~---

[Archiver]

Dieter,

The indicator syntax in the GAMS/Cplex option file is not as rich as
the GAMS syntax. The syntax is

indic equ(i,j)$bvar(i,j) 0 or 1

There can be only one variable (no expression) after the $ and the
variable has to be a binary variable.

Michael Bussieck - GAMSWorld Coordinator

On Jan 21, 7:28 am, "dieter.kunstm...@googlemail.com"
wrote:
> > Hello,
> > many Thanks an Ervin.
> >
> > can someone suggest how i can make for CPLEX.
> >
> > i write in GAMS
> > *--------------
> >
> > file fcplex Option file /cplex.opt/;
> > putclose fcplex 'indic Constr_1(c,o)$((Q(c,o)*(P(t,c,o)-Price(c))) ge
> > 0)'/
> > 'indic Price_2(c,o)$((Q_(c,o)*(P_u(t,c,o)-Price(c)) le
> > 0)';
> > mod1.optfile=1;
> >
> > loop((t,c,o),
> > put fcplex 'indic' Constr_1.tn(c,o)'$'x.tn(c,o) '$'P(c,o)'$'Q
> > (c,o)'$Price(c) yes/
> > 'indic' price_2.tn(t,c,o)'$$'x.tn(c,o) '$'P
> > (c,o)'$'Q(c,o)'$Price(c) yes/
> > );
> > putclose fcplex
> > Option miqcp=cplex;
> >
> > $offtext
> > solve mod1 using dnlp maximizing WElfare_global;
> >
> > *---------------------------------
> >
> > and i agree the X over all c and o =1. this is false...
> >
> > thanks
> > Dieter
--~--~---------~--~----~------------~-------~--~----~
To post to this group, send email to gamsworld@googlegroups.com
To unsubscribe from this group, send email to gamsworld+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/gamsworld?hl=en
-~----------~----~----~----~------~----~------~--~---