A Hybrid Methodology for Tackling the Irregular Strip Packing Problem*

This paper presents a hybrid approach for tackling the irregular strip packing problem, which requires a set of polygons to be placed within a rectangular object. The methodology includes a meta-heuristic engine (e.g., a genetic algorithm) that determines the order in which polygons are handled by a placement heuristic. In addition, differently from several approaches presented in the literature, we investigate the application of the no-fit polygon as a placement tool for obtaining local optima. The results are further improved by a shrinking algorithm that works within the meta-heuristic component. To assess the potentials of the proposed methodology, computational experiments performed on a set of difficult benchmark instances of the irregular strip packing problem are discussed here for evaluation purposes.