Nesting problems

• Ano de defesa: 2016
• Autor(a) principal: Cherri, Luiz Henrique
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: eng
• Instituição de defesa: Biblioteca Digitais de Teses e Dissertações da USP
• Palavras-chave em Português: Corte e empacotamento de peças irregulares; Ferramentas geométricas; Heurísticas; Modelos de programação inteira mista; Modelos de programação por restrições
• Link de acesso: http://www.teses.usp.br/teses/disponiveis/55/55134/tde-16112016-150256/

The two-dimensional irregular cutting and packing problems (aka nesting problems) have been studied over the past six decades and consist in cutting (packing) convex and non-convex small pieces from (in) large boards without overlapping. There are several variants of this problem that are defined according to the board shapes and the objective of each problem. There are a number of heuristics proposed in the literature to solve irregular cutting and packing problems, but only few mixed-integer programming models. Specifically, these models were developed for the irregular strip packing problem, that consists in packing pieces into a single board with fixed width and length to be minimized. For the other problem variants, there is no exact methods presented in the literature. The main difficulty in solving irregular cutting and packing problems is how to handle with the geometric constraints. These constraints depend on the type of placement of the pieces on the board that can be continuous or discrete. In this thesis, we present two mixed-integer programming models for the irregular strip packing problem in which the pieces can be continuously placed on the board. These models do not demand complex structures to be built. We also present a new dot data structure to store the information on the placement of the pieces and overlapping positions bringing flexibility and efficiency to discrete approaches. Using this structure, a matheuristic is proposed, combining the advantages of the models with discrete and continuous placement positions for the pieces on the board. Furthermore, constraint programming models for several variants of irregular cutting and packing problems are exploited. For some variants, these models are the first modelling representation. A new global constraint is developed to eliminate the overlap among pieces. Computational experiments were conducted to evaluate the developed approaches.