A growing collection of bilevel problems
mb_2007_01 : Linear-Linear problem from (Mitsos & Barton, 2007)
Objective values | Solution point |
---|---|
F* = 1.000 | y* = 1.000 |
f* = -1.000 |
The inner program has the unique optimal solution (y = 1) and therefore the bilevel program has the unique feasible point (see illustration on the right hand side) and therefore unique global minimum (see illustration on the left hand side). The inner program is linear and therefore the KKT conditions are both necessary and sufficient for a minimum.
Outer Problem | Inner Problem |
---|---|
![]() |
![]() |
Original source:
Other sources:
AMPL
formatvar y >= -1, <= 1; # Inner variable
var l{1..2} >= 0, <= 2; # KKT Multipliers
minimize outer_obj: y; # Outer objective
subject to
# Inner objective:
inner_obj: -y = 0;
# KKT conditions:
stationarity: -1 - l[1] + l[2] = 0;
complementarity_1: (y + 1)*l[1] = 0;
complementarity_2: (-y + 1)*l[2] = 0;