In rare instances the code generates an invalid GMIC. Currently I am just monitoring this as the cause is not apparent after checking for floating point and rounding issues. As long as it does not affect the size of the branch and bound tree, then we should be ok. Worst case, we can solve problems ahead of time and just toss cuts that would violate the optimal solution.
In rare instances the code generates an invalid GMIC. Currently I am just monitoring this as the cause is not apparent after checking for floating point and rounding issues. As long as it does not affect the size of the branch and bound tree, then we should be ok. Worst case, we can solve problems ahead of time and just toss cuts that would violate the optimal solution.