A growing collection of bilevel problems
Linear-Quadratic bilevel problem from [Mitsos & Barton, 2007]
Objective values | Solution point(s) |
---|---|
F* = 1.000 | y* = 1.000 |
f* = -1.000 |
The inner program has the unique optimal solution y* = 1 and therefore the bilevel problem has the unique feasible point y = 1.
Outer Problem | Inner Problem |
---|---|
![]() |
![]() |
AMPL
formatvar y >= -0.5, <= 1; # Inner variables
var l{1..2} >= 0, <= 2; # KKT Multipliers
minimize outer_obj: y; # Outer objective
subject to
# Inner objective:
inner_obj: -y^2 = 0;
# KKT conditions:
stationarity: -2*y - l[1] + l[2] = 0;
complementarity_1: l[1]*(-0.5 - y) = 0;
complementarity_2: l[2]*(y - 1) = 0;