Changes between Version 1 and Version 2 of Ticket #29338, comment 9


Ignore:
Timestamp:
06/24/20 13:46:18 (2 years ago)
Author:
slabbe
Comment:

Legend:

Unmodified
Added
Removed
Modified
  • Ticket #29338, comment 9

    v1 v2  
    1 I made that choice to make it explicit that we are not using the dlx solver (we can not compare timing for example, because of the initialization of the rows is already done when using "dlx") and to make explicit that there is a overhead in doing the translation of the thousands of rows into something else.
     1I made that choice to make it explicit that we are not using the dlx solver (we can not compare timing (at least on the first call) for example, because of the initialization of the rows is already done when using "dlx") and to make explicit that there is a overhead in doing the translation of the thousands of rows into something else.
    22
    33I think `one_solution(solver='Gurobi')` should not be something so much desirable, since if this is really what is the most efficient, then one should just construct the `MILP` directly.