Nesting problems

Autor(es)

Luiz Henrique Cherri

• Tipo de documento: Tese de Doutorado
• Imprenta: São Carlos.
• Instituição: Universidade de São Paulo (USP). Instituto de Ciências Matemáticas e de Computação (ICMC/SB)
• Data de defesa: 2016-05-13

Membros da banca

Franklina Maria Bragion de Toledo; Flávio Keidi Miyazawa; Ramon Alvarez valdes Olaguibel; Maria Cristina de Carvalho Alves Ribeiro; Horacio Hideki Yanasse

• Orientador: Franklina Maria Bragion de Toledo; Maria Antónia da Silva Lopes de Carravilla

Resumo

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. (AU)

• Processo FAPESP: 12/18653-8 - O problema de corte de peças irregulares
• Beneficiário: Luiz Henrique Cherri
• Modalidade de apoio: Bolsas no Brasil - Doutorado