Description

Test problem 9.2.4 in Handbook of Test Problems in Local and Global Optimization
Test problem 9.1.3 on http://titan.princeton.edu/TestProblems/chapter9.html

References:

Floudas, C A, Pardalos, P M, Adjiman, C S, Esposito, W R, Gumus, Z H, Harding,
S T, Klepeis, J L, Meyer, C A, and Schweiger, C A, Handbook of Test Problems in
Local and Global Optimization. Kluwer Academic Publishers, 1999.

Candler, W, and Townsley, R, A Linear Two-Level Programming Problem.
Comp. Op. Res.9 (1982), 59-76.

Contributor: Alex Meeraus and Jan-H. Jagla, December 2009

Small Model of Type : BP

Category : GAMS EMP library

Main file : flds913.gms

$title Princeton Bilevel Optimization Example 9.1.3 (FLDS913,SEQ=29)

$onText

  Test problem 9.2.4 in Handbook of Test Problems in Local and Global Optimization
  Test problem 9.1.3 on http://titan.princeton.edu/TestProblems/chapter9.html

References:

Floudas, C A, Pardalos, P M, Adjiman, C S, Esposito, W R, Gumus, Z H, Harding,
S T, Klepeis, J L, Meyer, C A, and Schweiger, C A, Handbook of Test Problems in
Local and Global Optimization. Kluwer Academic Publishers, 1999.

Candler, W, and Townsley, R, A Linear Two-Level Programming Problem.
Comp. Op. Res.9 (1982), 59-76.

Contributor: Alex Meeraus and Jan-H. Jagla, December 2009

$offText

*Solution of problem 9.1.3 on the web
scalar x1_l /  0   /
       x2_l /  0.9 /
       y1_l /  0   /
       y2_l /  0.6 /
       y3_l /  0.4 /
       tol  / 1e-6 /;

Variables z, z_in; Positive Variables x1, x2, y1, y2, y3;
equations ob, ob_in, c2, c3, c4;

ob..     - 8*x1 - 4*x2 + 4*y1 - 40*y2 -   4*y3 =e= z;

ob_in..      x1 + 2*x2 +   y1 +    y2 +   2*y3 =e= z_in;
c2..                   -   y1 +    y2 +     y3 =l= 1;
c3..       2*x1        -   y1 +  2*y2 - 0.5*y3 =l= 1;
c4..              2*x2 + 2*y1 -    y2 - 0.5*y3 =l= 1;

Model bilevel / all /;

$echo bilevel x1 x2 min z_in y1 y2 y3 ob_in c2 c3 c4 > "%emp.info%"

* Start from reported solution
x1.l = x1_l;
x2.l = x2_l;
y1.l = y1_l;
y2.l = y2_l;
y3.l = y3_l;
* The reported solution is hardly a global attractor: fix x2 to give the solver some help,
* otherwise the solution found depends on the solver used
x2.fx = x2_L;

solve bilevel using emp minimizing z;

abort$(   (abs(x1.l - x1_l) > tol)
       or (abs(x2.l - x2_l) > tol)
       or (abs(y1.l - y1_l) > tol)
       or (abs(y2.l - y2_l) > tol)
       or (abs(y3.l - y3_l) > tol) ) 'Deviated from reported solution';


* and now verify that we get the solution after unfixing x2
x2.up = INF;

solve bilevel using emp minimizing z;

abort$(   (abs(x1.l - x1_l) > tol)
       or (abs(x2.l - x2_l) > tol)
       or (abs(y1.l - y1_l) > tol)
       or (abs(y2.l - y2_l) > tol)
       or (abs(y3.l - y3_l) > tol) ) 'Deviated from reported solution';