Abstract

In this work, we test the hypothesis that heuristic preprocessing can be used to increase the computational efficiency of branch and cut solutions to mine sequencing problems. Three preprocessing algorithms based on problem-specific cutting planes were developed and used to investigate the effect of heuristic preprocessing on computational efficiency. The cutting planes were based on: (i) the greedy algorithm; (ii) a strategy to mine sections with no precedence constraints first; and (iii) a strategy to mine sections in the mine’s development area first. A sample room and pillar problem modeled as binary integer linear programming was used to investigate computation efficiency of using the cutting planes for preprocessing. Of the three, the best cutting plane to preprocess the sample problem was the one based on mining the sections in the development area first. The execution of the preprocessing algorithm took less than 1% of the overall solution time prior to preprocessing. Preprocessing the problem resulted in 29% less computational time and 43% fewer number of iterations.