Our aim from the very beginning is to solve an integer linear program. Yet, all that we have algorithmically accomplished so far is to solve the linear relaxation of a Dantzig-Wolfe reformulation. When the solution is fractional, we need to cut it, or branch. From a bird’s eye view, all that remains to do is to use the column generation algorithm to solve the relaxation, and possibly strengthen it by additional valid inequalities, in every node of the branch-and-bound tree. Zooming in, we see that some subtleties have to be respected or can even be exploited.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Branch-Price-and-Cut

  • Jacques Desrosiers,
  • Marco Lübbecke,
  • Guy Desaulniers,
  • Jean Bertrand Gauthier

摘要

Our aim from the very beginning is to solve an integer linear program. Yet, all that we have algorithmically accomplished so far is to solve the linear relaxation of a Dantzig-Wolfe reformulation. When the solution is fractional, we need to cut it, or branch. From a bird’s eye view, all that remains to do is to use the column generation algorithm to solve the relaxation, and possibly strengthen it by additional valid inequalities, in every node of the branch-and-bound tree. Zooming in, we see that some subtleties have to be respected or can even be exploited.