How to use EMP for getting KKT?

GAMS
1

Hi,

I want to use JAMS solver to use emp for getting KKT of a linear model.
I have read the EMP manual but I did not understand how to interpret the MCP from result. Is there any special file which those equation is created?
I used the example of EMP manual as follow:

variables f,z; positive variables x,y;
equations g, h, defobj;

g… x + y =l= 1;
h… x + y - z =e= 2;
defobj… f =e= -3*x + y;

model comp / defobj, g, h /;

file info / ‘%emp.info%’ /;
putclose info / ‘modeltype mcp’;

solve comp using emp minimizing f;

— The solution report and solvar are as follow:

S O L V E S U M M A R Y

MODEL comp OBJECTIVE f
TYPE EMP DIRECTION MINIMIZE
SOLVER JAMS FROM LINE 13

**** SOLVER STATUS 1 Normal Completion
**** MODEL STATUS 1 Optimal
**** OBJECTIVE VALUE -3.0000

RESOURCE USAGE, LIMIT 0.183 1000.000
ITERATION COUNT, LIMIT 0 2000000000
EVALUATION ERRORS 0 0

JAMS 1.0 Dec 13, 2010 23.6.3 WEX 22848.22869 WEI x86_64/MS Windows


JAMS - Solver for Extended Mathematical Programs (EMP)

— EMP Summary (errors=0)
Adjusted Constraint = 0
Flipped Constraints = 0
Dual Variable Maps = 0
Dual Equation Maps = 0
VI Functions = 0
Equilibrium Agent = 0
Bilevel Followers = 0
Disjunctions = 0
Logical Constraints = 0
— The model Z:\home2\myazdani\Documents\gamsdir\projdir\225a\emp.dat will be s
olved by GAMS

— Returning from GAMS step

PATH Dec 13, 2010 23.6.3 WEX 22848.22869 WEI x86_64/MS Windows

5 row/cols, 10 non-zeros, 40.00% dense.

Path 4.7.02 (Fri Feb 11 13:11:19 2011)
Written by Todd Munson, Steven Dirkse, and Michael Ferris

— Equations skipped = 1 Equations added = 3
— Variables skipped = 1 Variables added = 2

LOWER LEVEL UPPER MARGINAL

---- EQU defobj . . . 1.000
---- EQU g -INF 1.000 1.000 -3.000
---- EQU h 2.000 2.000 2.000 .

LOWER LEVEL UPPER MARGINAL

---- VAR f -INF -3.000 +INF .
---- VAR z -INF -1.000 +INF .
---- VAR x . 1.000 +INF .
---- VAR y . . +INF 4.000


**** REPORT SUMMARY : 0 NONOPT
0 INFEASIBLE
0 UNBOUNDED
0 ERRORS

**** REPORT FILE SUMMARY

info Z:\home2\myazdani\Documents\gamsdir\projdir\225a\empinfo.dat


EXECUTION TIME = 0.031 SECONDS 2 Mb WEX236-236 Feb 11, 2011



**** FILE SUMMARY

Input \eead.ee.kth.se\dfs\home2\myazdani\Documents\gamsdir\projdir\Untitle
d_13.gms
Output \eead.ee.kth.se\dfs\home2\myazdani\Documents\gamsdir\projdir\Untitle
d_13.lst

–
To view this discussion on the web visit https://groups.google.com/d/msg/gamsworld/-/l5PjT21z6ZEJ.
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.

2

The JAMS solver creates the KKT of the nlp that you specify, and this is an MCP.
This MCP is formed (at the scalar level) and passed onto the PATH solver that generates the solution.

By default, you do not see the MCP model that is generated, it is simply solved and the results given back into your original gams model. There are options that allow you to see the generated model, but these are probably not in the form you would like (they are mapped into a scalar model and done equation by equation). However, if that is what you want, the option file for jams (jams.opt) would be something like:

filename mcpmod.gms
dict dict.txt

(Two files will be created - mcpmod.gms contains the mcp model, while dict.txt would show the mapping between the names in your original model and the scalar names in the mcp model).

Hope this helps.

Michael Ferris

On Sep 5, 2012, at 9:59 AM, Mohammad Yazdani wrote:

Hi,

I want to use JAMS solver to use emp for getting KKT of a linear model.
I have read the EMP manual but I did not understand how to interpret the MCP from result. Is there any special file which those equation is created?
I used the example of EMP manual as follow:

variables f,z; positive variables x,y;
equations g, h, defobj;

g… x + y =l= 1;
h… x + y - z =e= 2;
defobj… f =e= -3*x + y;

model comp / defobj, g, h /;

file info / ‘%emp.info%’ /;
putclose info / ‘modeltype mcp’;

solve comp using emp minimizing f;

— The solution report and solvar are as follow:

   S O L V E      S U M M A R Y

 MODEL   comp                OBJECTIVE  f
 TYPE    EMP                 DIRECTION  MINIMIZE
 SOLVER  JAMS                FROM LINE  13

**** SOLVER STATUS 1 Normal Completion
**** MODEL STATUS 1 Optimal
**** OBJECTIVE VALUE -3.0000

RESOURCE USAGE, LIMIT 0.183 1000.000
ITERATION COUNT, LIMIT 0 2000000000
EVALUATION ERRORS 0 0

JAMS 1.0 Dec 13, 2010 23.6.3 WEX 22848.22869 WEI x86_64/MS Windows

JAMS - Solver for Extended Mathematical Programs (EMP)

— EMP Summary (errors=0)
Adjusted Constraint = 0
Flipped Constraints = 0
Dual Variable Maps = 0
Dual Equation Maps = 0
VI Functions = 0
Equilibrium Agent = 0
Bilevel Followers = 0
Disjunctions = 0
Logical Constraints = 0
— The model Z:\home2\myazdani\Documents\gamsdir\projdir\225a\emp.dat will be s
olved by GAMS

— Returning from GAMS step

PATH Dec 13, 2010 23.6.3 WEX 22848.22869 WEI x86_64/MS Windows

5 row/cols, 10 non-zeros, 40.00% dense.

Path 4.7.02 (Fri Feb 11 13:11:19 2011)
Written by Todd Munson, Steven Dirkse, and Michael Ferris

— Equations skipped = 1 Equations added = 3
— Variables skipped = 1 Variables added = 2

                   LOWER     LEVEL     UPPER    MARGINAL

---- EQU defobj . . . 1.000
---- EQU g -INF 1.000 1.000 -3.000
---- EQU h 2.000 2.000 2.000 .

                   LOWER     LEVEL     UPPER    MARGINAL

---- VAR f -INF -3.000 +INF .
---- VAR z -INF -1.000 +INF .
---- VAR x . 1.000 +INF .
---- VAR y . . +INF 4.000

**** REPORT SUMMARY : 0 NONOPT
0 INFEASIBLE
0 UNBOUNDED
0 ERRORS

**** REPORT FILE SUMMARY

info Z:\home2\myazdani\Documents\gamsdir\projdir\225a\empinfo.dat

EXECUTION TIME = 0.031 SECONDS 2 Mb WEX236-236 Feb 11, 2011

**** FILE SUMMARY

Input \eead.ee.kth.se\dfs\home2\myazdani\Documents\gamsdir\projdir\Untitle
d_13.gms
Output \eead.ee.kth.se\dfs\home2\myazdani\Documents\gamsdir\projdir\Untitle
d_13.lst

–
To view this discussion on the web visit https://groups.google.com/d/msg/gamsworld/-/l5PjT21z6ZEJ.
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.

\