An approach for the cutting stock problem of hardboard sheets

In this paper we present an appproach for the cutting stock problem of a Brazilian hardboard enterprise. The problem consists of determining the best patterns to be cut by a programmable samba saw, composed of a set of circular saws for longitudinal and transversal cuts. The samba saw involves unusual constraints. These include a limitation on the number of item types and on the difference between the largest and smallest length of the types in the cutting pattern. There are also ‘usual’ constraints such as the availability of longitudinal and transversal saws, and orthogonal and guillotine cuttings in at most two stages without trimming. The problem is modeled in two phases, each one as an integer program, and solved utilizing an implicit enumeration method. The computational performance of the approach is illustrated solving hundreds of examples.