Dimensionamento e seqüenciamento de lotes de produção na indústria de suplementos para nutrição animal
Toso, Eli Angela Vitor
MetadataShow full item record
This work studies the integrated lot sizing and scheduling problem in the animal feed compound industry. The lot sizing problem in this industry consists of deciding which and how much to produce in each period, in order to minimize overtime and storage costs. The sequencing problem consists of sequencing the production lots, in order to minimize the setups (that eat into the available capacity), and to avoid the risks of residual contamination. The main difference of this problem in relation to the ones in literature is the structure of the setup times. Using a case study in a company of the sector, four approaches are proposed to model and solve the problem. The first two are based on the General Lot Sizing and Scheduling Problem (GLSP) with sequence dependent setup times. The other two approaches consist of a reformulation of the GLSP model, considering the lot sequencing as an Asymetric Travelling Salesman Problem (ATSP). Either modeling approach GLSP and ATSP is proposed for two company strategies related to the cleaning of the production line, called (1) Independent Sequences , where it is assumed that at the end of each period a complete cleaning in the production line is carried out; and (2) Dependent Sequences , where the sequence at the beginning of each period depends on the preparation state of the line in the previous period (setup carryover). The model GLSP Independent Sequences is solved by the branch-and-cut method (using the software AMPL/CPLEX), with limited computational time. To solve the model GLSP Dependent Sequences , besides the branch-and-cut method, two heuristic relax-and-fix procedures are proposed . To solve the model ATSP Independent Sequences the subtour elimination method is used. In the case of the model ATSP Dependent Sequences , as well as the subtour elimination method, the patching subtours method is used. According to experiments carried out with real data, the models and methods proposed solve the problem satisfactorily, getting better results that the company. Of the different approaches proposed, the most appropriate for the problem appears to be the reformulation ATSP with the patching method and the strategy Dependent Sequences .