A heuristic method for generating constrained two-dimensional guillotine cutting patterns based on the exact method by Christofides and Hadjiconstantinou (1995) was presented in Silveira and Morabito (2002). The method combines a state space relaxation of a dynamic programming formulation, a subgradient optimization procedure and an inner heuristic that turn infeasible solutions provided in each step of the optimization procedure into feasible solutions. In this work, the method of Silveira and Morabito is modified by using a more effective inner heuristic, and an and/or-graph search approach in order to generate good initial solutions. Results for benchmark and randomly generated instances indicate that the refined method's performance is superior to the previous one, and it is competitive face to other methods proposed in the literature.
Cutting problems; constrained two-dimensional guillotine cutting patterns; dynamic programming; heuristics
