In this paper we study a particular case of two-dimensional cutting problems named constrained guillotine cutting (CGC). The CGC is an NP-hard problem that appears in different industrial processes of cutting rectangular plates, such as in the glass and circuit board industries. To solve the problem we present a variation of the exact method of CHRISTOFIDES & HADJICONSTANTINOU (1995), based on a state space relaxation of a dynamic programming formulation of the CGC, a procedure of subgradient optimization type, and a feasibility heuristic. The result is a method without guarantee of optimality, however, faster and able to solve larger problems than the exact method of Christofides and Hadjiconstantinou. The computational performance of the approach is evaluated solving several examples of the literature as well as randomly generated examples, and comparing the solutions obtained with the ones of Christofides and Hadjiconstantinou’s method and the well-known heuristic of WANG (1983).
cutting problem; constrained two-dimensional cutting patterns; dynamic programming; heuristics
