A Pre-estimate Approach to Find a Good Initial Point to Integer Linear Programming Problems

Abstract

Abstract-Integer Linear Programming problems are significant in many areas, and there exist many classic methods to deal with such problems. Based on the intuition that a proper pre-estimation of those problems may help find a good initial point that is relatively closed to the final solution, we constructed a framework to do such pre-estimate by introducing . We proposed two methods to estimate by trials and by direct estimation. In a case study our method reduces about 7% of computational expense compared to the Branch and Bound Algorithm.