An and-or-graph approach for two-dimensional cutting problems

The problem of generating guillotine cutting patterns for a rectangular plate is studied and a type of structure is proposed for representing the solution of the problem, called and-or graph, which is commonly used in the Artificial Intelligence environment. To search the graph we combined two classical strategies: depth-first and hill-climbing. Further, some heuristics are considered and computational results are presented, relating their performance on examples both from literature as well as randomly generated.