Sistema de apoio à decisão para programação da produção em fundições de mercado.
Abstract
The aim of production scheduling is to time the use of resources to carry
out the production orders required to meet a previously defined production need. Due to
the changes in the production settings, defined by the increase in competition and
diversity of products, there has been increasing complexity in the development of
production scheduling process, mainly when complex production settings are
considered.
Despite the developments in the area, a significant number of production
settings still do not use the applicable solutions in the area. The initial aims of this work
were to create alternatives to fill this gap. From these findings and based on an initial
survey, carried out in the interior of São Paulo State, regarding the productive settings
of foundries which work with customized products made to order, also known as steel
casting foundries, effective alternatives for the production scheduling process for this
specific production settings were sought.
This work presents a viable proposal for the production scheduling
process in steel casting foundries, considering its main particularities and restrictions,
using a decision support system, which employs a single cycle system known as Period
Batch Control (PBC).
Among the results obtained, a methodology for the rationalization and
systematization of the decision-making process in the production control was proposed.
For the development of a decision support system in ordering process, the following
viable decision-making models were proposed: a model of binary integer programming;
a model based on classical approximation methods, specifically the method known as
beam search ; and a meta-heuristic model based on genetic algorithms. The results
obtained from the application of decision-making models in experimental production
situations similar to the reality of production in steel casting foundries have confirmed
the practical viability of these models, providing a comparison of the results between
distinct modeling techniques. It has also contributed to the designing of guidelines for
other applications of decision support systems in the operational area of complex
production settings. Finally, the suitable of the proposed system was conformed by a
later survey carried out in the main steel casting foundries in the interior of São Paulo
State.