Abstract

Makespan, is a differentiating factor in productive enterprises, as its optimization contributes to operational efficiency. This study analyzes a real case of dryer loads of wood in green state in a sawmill in Chile. The study considers a set of 10 parallel machines with times of set-up independent of sequence, with 161 jobs, on a horizon of monthly planning. The proposed model of integer programming is ejecuted in AMPL and a statistical analysis is realized for solution quality and computing time with commercial solvers CPLEX and GUROBI. The results of the computational experiment allow to reduce the makespan by 8.5% with respect to the company’s current scheduling approach. The CPLEX solver, performs better in 59.3% of the cases analyzed. The most influential parameters for computing time were: in GUROBI cuts (evaluated at 0), in CPLEX mipcuts (evaluated at 2) and repeatpresolve (evaluated at 0). The latter difference of this parameter is statistically significant.