Hello,
The paper says the best-first strategy is applied, with which the node whose parent has the highest lower bound (father_lb) will be explored first. However, the priority_queue in the code is a min-heap in which the node with smaller father_lb has higher priority. So it seems they are inconsistent? Did I miss something?
This is also verified by BBTree::update_lb() in which the top element in priority_queue is used to update the global lower bound. This is valid in the min-heap case.
Thanks you!
Hello,
The paper says the best-first strategy is applied, with which the node whose parent has the highest lower bound (
father_lb
) will be explored first. However, the priority_queue in the code is a min-heap in which the node with smallerfather_lb
has higher priority. So it seems they are inconsistent? Did I miss something?This is also verified by
BBTree::update_lb()
in which the top element in priority_queue is used to update the global lower bound. This is valid in the min-heap case. Thanks you!