In this work we review some linear and nonlinear integer models to generate two stage two-dimensional guillotine cutting patterns, including the constrained, non constrained, exact and non exact cases. These problems are particular cases of the two dimensional knapsack problems. We also present new models to generate these cutting patterns, based on adaptations and extensions of models that generate one-group constrained two dimensional cutting patterns. Two stage patterns arise in different cutting processes like, for instance, in the furniture industry and wooden hardboards. The models are useful for the research and development of more efficient methods, exploring particular structures, the model decomposition, model relaxations etc. They are also useful to evaluate the performance of heuristics, since they allow (at least for problems of moderate sizes) an estimative of the optimality gap of the solutions obtained by heuristics. To illustrate the application of the models we analyze the results of some computational experiments with instances of the literature and other generated randomly. The results were produced using a known commercial software and they show that the necessary computational effort to solve the models can be very different.
Cutting and packing problems; Two-dimensional knapsack; Two-stage guillotine cut; Linear and nonlinear integer models; Furniture industry
