On Virtual Development of Manufacturing Systems -Proposal for a Modular Discrete Event Simulation Methodology

Doktorsavhandling, 2006

Today’s production is characterized by frequent changes and uncertainty due to short product life cycles. Outsourcing to low-cost countries is a common way to solve flexibility and cost problems. Swedish manufacturing companies need to use knowledge as an advantage in order to compete successfully. This thesis formulates one approach on how to reuse and take advantage of knowledge and data for discrete event simulation of manufacturing systems. The approach uses modularity and division of knowledge in order to present a methodology for modular discrete event simulation. The objective of the research building this thesis is to formulate the demands on a methodology for modular discrete event simulation, which is increasing the reuse of data and knowledge-intensive tasks. This in turn enables sharper focus on the valueadding performance improvement, for which a discrete event simulation study can generate a firm basis. In order to meet this objective, research has been conducted in the following steps:

· Prerequisite studies in order to find the research gap in case studies and papers
· Confirmation of the research gap and its current status in case studies and papers
· Management of the research gap by formulation of a methodology in this thesis
· Validation of the methodology in four case studies

The proposed methodology covers activities for simulation specialists, in terms of module building, simulation users in terms of model building and decision support, and simulation observers, in terms of decision support and visualization. The methodology is based on traditional discrete event simulation methodologies, with additional effort put into division of knowledge-intensive tasks in order to streamline the model building and the reuse of data and knowledge. This is achieved through modularization and parameterization of the building blocks used in discrete event
simulation model building with module boundaries similar to those in the real-world manufacturing system. It is concluded that the modular discrete event simulation methodology can provide
profitable advantages for OEM suppliers of manufacturing systems and their nearby interactors. Important parts of the methodology have been validated in industrial case studies; however, general benefits for any manufacturing system were not found, only for OEM suppliers of manufacturing equipment and those with direct relations to these kinds of systems.