Advanced search
Start date

A general integrated lot-sizing and cutting stock problem: mathematical modelling and solution methods

Full text
Gislaine Mara Melega
Total Authors: 1
Document type: Doctoral Thesis
Press: São José do Rio Preto. 2017-03-29.
Institution: Universidade Estadual Paulista (Unesp). Instituto de Biociências Letras e Ciências Exatas. São José do Rio Preto
Defense date:
Advisor: Silvio Alexandre de Araujo

In this thesis, the subject of interest is in treating, in an integrated way, two wellknown problems in the literature. This integration is referred in the literature as the integrated lot-sizing and cutting stock problem. The basic idea is to consider, simultaneously, the decisions related to both problems so as to capture the interdependency between these decisions in order to obtain a better global solution. We propose a mathematical model for a general integrated lot-sizing and cutting stock (GILSCS) problem. This model considers multiple dimensions of integration and enables us to classify the current literature, in terms of mathematical models, in this field. The main classification of the literature is organized around two types of integration. In a planning horizon which consists of multiple periods, the inventory provides a link between the periods. This integration across time periods constitutes the first type of integration. The general problem also considers the production in different levels: objects are fabricated or purchased and then, they are cut to produce the pieces which are then assembled as components in the production of final products. The integration between these production levels constitutes the second type of integration. The literature review also enables us to point out interesting areas for future research. The behavior of a solution to this type of problem, with three levels of production and several time periods, is studied considering the development of solution approaches that overcome the difficulties of the problem, which are the high number of cutting patterns, multi-level structures and the binary values of the setup variables. The solution methods proposed to the GILSCS problem are based on two known strategies from the literature which are used successfully to solve the problems separately, which are the column generation procedure and decomposition heuristics based on relax-and-fix procedure. These strategies and their variations are combined into an optimization package in a computational study with randomly generated data. A literature review, in terms of solution methods, to the integrated problem, is also presented. Other contributions of this thesis consist of proposing different mathematical models for the integrated problem combining alternative models for each one of the problems separately. In this study, the aim is to compare and evaluate, with an extensive computational study, the quality and the impact of these dfifferent formulations. Another study is an application of the integrated problem in a small furniture factory, in which specific constraints related to the industrial environment are addressed, such as, safety stock level constraints and saw cycles constraints. The solution obtained from the proposed model is compared to a simulation of the common practice in the company. (AU)

FAPESP's process: 12/20631-2 - The Integrated Lot Sizing and Cutting Stock Problem
Grantee:Gislaine Mara Melega
Support type: Scholarships in Brazil - Doctorate