[gecode-users] Propagating to Fixpoint

[Max Ostrowski]

Can the order of propagation influence the final result?


Given a list of reified constraints L posted with ICL_DEF

for (i in L)
{
i == true;
status();
}

// the space is failed

Now the same thing with the list sorted by some order

randomize_order(L)
for (i in L)
{
i == true;
status();
}

// the space not failed



Can this happen (not using circuit and nvalues) only usign linear
constraints.
Why does this happen.


Best,
Max