Improving the LP bound of a MILP by dual concurrent branching and the relationship to cut generation methods

In this paper branching for attacking MILP is investigated. Under certain circumstances branches can be done concurrently. By introducing a new calculus it is shown there are restrictions for dual values. As a second result of this study a new class of cuts for MILP is found, which are defined by those values. This class is a superclass of all other classes of cuts. Furthermore the restrictions of the dual values can be used for studying the addition of arbitrary inequalities. This theory has similarities but also big differences to the theory of disjunctive programming.

Citation

Report is to be published. Comments and supporters are strongly welcomed for this!

Article

Download

View PDF