Decomposition Branching for Mixed Integer Programming

Applications of mixed integer programming can be found in many industries, such as transportation, healthcare, energy, and finance, their economic impact is significant. It also well known that programs (MIPs) very difficult to solve. Their challenge continues stimulate research the design implementation efficient effective techniques better solve them. In this study, we introduce a novel powerful approach for solving certain classes (MIPs): decomposition branching. Two seminal widely used MIPs, branch-and-bound decomposition, form its foundation. Computational experiments with instances weighted set covering problem regionalized p-median facility location assignment range constraints demonstrate efficacy: it explores far fewer nodes orders magnitude faster than commercial solver an automatic Dantzig-Wolfe approach.