Solving realistic cutting stock problems using metaheuristics

摘要

The paper considers a realistic pattern-oriented Cutting Stock Problem identified in a national manufacturer of metal profile, where it has to be selected the best combination of cutting patterns in order to satisfy customer demands with material available in stock, looking for minimizing both trim loss and cutting times. Two different heuristic procedures based on genetic algorithms and particle swarm optimization are developed in order to solve the auxiliary problem arising from the LP formulation of the cutting stock problem which consists on the generation of those matrixes which contain the information regarding cutting patterns.