THESIS
1999
xi, 164 leaves : ill. ; 30 cm
Abstract
This dissertation is about location selection of manufacturing systems. The strategic decision issue regarding long-term resources commitment, commonly known as location theory, has been studied over ninety years. Traditional manufacturing location theory focuses essentially on locational decision issues -- where to make, giving product decisions -- what to make and manufacturing process decisions -- how to make as premises....[
Read more ]
This dissertation is about location selection of manufacturing systems. The strategic decision issue regarding long-term resources commitment, commonly known as location theory, has been studied over ninety years. Traditional manufacturing location theory focuses essentially on locational decision issues -- where to make, giving product decisions -- what to make and manufacturing process decisions -- how to make as premises.
In this thesis, our primary focus will be on the Location Problems with Product and Process Considerations (or equivalently the Product-Process-Location Problem, abbreviated as PPLP). We attempt to address simultaneously the product, manufacturing process, and manufacturing location selection issues in order to satisfy customers best, and to leverage the capabilities and resources on a worldwide basis.
Because of the interlocking of selecting product, manufacturing process, and manufacturing location, the formulation of PPLP becomes the first challenge of this research. Based on the axiomatic design theory, a domain-system based framework is proposed to define the PPLP and its solution space. The construction of PPLP involves the participation of individuals from different fields. Due to the emphasis on different facets by different participants, PPLP becomes typically an over-constrained problem with a large number of decision variables, where the constraints appear in various mathematical forms. In addition, PPLP needs to consider multiple objectives because a high return on investment, a good product selection, and a good process selection have to be concurrently realized. These features pose some other challenges for representing and solving PPLP.
In this thesis, PPLP is modeled as a dynamic constraint satisfaction problem. It is further represented by a structured constraint network and corresponding constraint matrix. An enhanced constraint management approach is then developed to evaluate the constraint network for finding a set of satisfactory solutions. The main components of this enhanced constraint management approach are as follows: problem partition, consistency checking, backtrack-free search, constraint propagation, evaluation and resolution.
As for the multiple objectives encountered in the product and process selection, an Analytic Hierarchy Process approach based on the probability of satisfying functional requirements is proposed. Two different approaches are compared. Some guidelines for using this approach are also presented.
A prototype system has been developed on the KAPPA-PC platform to implement the proposed representation schemes and solution methodology. The feasibility of the system is then illustrated by an example.
Post a Comment