[Bug-glpk] Wrong dual value using interior method

[xypron]

The interior solver outputs incorrect dual values
for a feasable solution:

var x;
var y;
maximize obj: x + y;
s.t. c1: 0.1 * y + x <= 1.;
s.t. c2: 0.1 * y + x >= 1.;
s.t. d1: y + 0.1 * x <= 5.;
s.t. d2: y + 0.1 * x >= 5.;
solve;
printf "c1: value = %d, dual = %f\n", c1.val, c1.dual;
printf "c2: value = %d, dual = %f\n", c2.val, c2.dual;
printf "d1: value = %d, dual = %f\n", d1.val, d1.dual;
printf "d2: value = %d, dual = %f\n", d2.val, d2.dual;
end;

./glpsol --interior -m test.mod
outputs
c1: value = 1, dual = -0.432591
c2: value = 1, dual = 1.341682
d1: value = 5, dual = -0.447088
d2: value = 5, dual = 1.356179

Best regards

Xypron
-- 
View this message in context: 
http://old.nabble.com/Wrong-dual-value-using-interior-method-tp30057475p30057475.html
Sent from the Gnu - GLPK - Bugs mailing list archive at Nabble.com.