Full text
| |
| Author(s): |
Marcos Mansano Furlan
Total Authors: 1
|
| Document type: | Doctoral Thesis |
| Press: | São Carlos. |
| Institution: | Universidade de São Paulo (USP). Instituto de Ciências Matemáticas e de Computação (ICMC/SB) |
| Defense date: | 2015-12-10 |
| Examining board members: |
Reinaldo Morabito Neto;
Silvio Alexandre de Araujo;
Marcos Nereu Arenales;
Bernardo Sobrinho Simões de Almada Lobo;
Debora Pretti Ronconi
|
| Advisor: | Reinaldo Morabito Neto; Maristela Oliveira dos Santos |
| Abstract | |
The pulp and paper industry has been increasing the commercial importance in recent decades due to the constant growing demand. The increasing competitiveness of this sector generated by the globalized economy and the difficulty to develop good production plans in complex production environments have motivated the search for new and effective decision support systems. Given these difficulties, in this thesis we address the lot sizing and scheduling problem focused on integrated pulp and paper mills. This is a problem of medium to short term planning, generally more focused on the short term as it covers detailed production schedules in planning horizons which do not exceed 30 days. In these integrated pulp and paper process the production decisions of digester, evaporator, recovery boiler and multiple paper machines are considered, apart from the inventory control of intermediate and final products. Mathematical models known in the literature were modified and extended to incorporate additional features of the problem, such as processes with multiple paper machines. In addition, constructive and improvement heuristics, and hybrid methods based on genetic algorithms combined with a commercial solver were developed, as well as combinations of these solution approaches. The methods developed were computationally tested and the best combinations of methods were defined. Overall, the results of these methods were superior to the solutions obtained by pure commercial solvers. Moreover, the alternative variation proposed of the improvement heuristic fix-and-optimize with exchanges in the objective function surpassed the other methods, obtaining the best results, regardless of the quality of the initial solution used. The main contribution of this thesis are the presentation of mathematical models that appropriately represents the problem under study, and the development of effective solution methods to deal with the problem. (AU) | |
| FAPESP's process: | 11/00274-8 - The integrated lotsizing and scheduling problem in the pulp and paper mill |
| Grantee: | Marcos Mansano Furlan |
| Support Opportunities: | Scholarships in Brazil - Doctorate |