Fix bin packing initial propagation #1113
Conversation
solver/src/main/java/org/chocosolver/solver/constraints/nary/binPacking/PropBinPacking.java
Outdated
Show resolved
Hide resolved
… in PropBinPacking
| @Override | ||
| public void propagate(int evtmask) throws ContradictionException { | ||
| if(PropagatorEventType.isFullPropagation(evtmask)) { | ||
| for(int j = 0; j<nbAvailableBins; j++) { |
There was a problem hiding this comment.
to be 100% safe, P[j] and R[j] should be cleared here I think
There was a problem hiding this comment.
I agree, reifying this constraint could lead to unexpected results if P[j] and R[j] are not cleared.
There was a problem hiding this comment.
Yes, I agree it would be safer (and it is now done)
solver/src/main/java/org/chocosolver/solver/constraints/nary/binPacking/PropBinPacking.java
Outdated
Show resolved
Hide resolved
cprudhom
left a comment
cprudhom
left a comment
There was a problem hiding this comment.
I agree with the correction, but it seems safer not to clean up P and R in full propagation.
mergify
bot
dismissed
cprudhom’s stale review
Pull request has been modified.
ArthurGodet
force-pushed
the
fix/binPacking
branch
from
db56caf to
5e6ab19
Compare
|
You mean, it would be safer to clean R and P in full propagation ? Or not clean them (as it is now) ? |
To set them to 0 as the first instruction of the |
* #1112 : fix bin packing initial propagation * move code initialising P and sumP from constructor to fullPropagation in PropBinPacking * clear R and P sets before full propagation in PropBinPacking
Fix bin packing's propagator initial propagation (described in #1112)