|
19th EMSS, 2007 |
|
19th European Modeling and Simulation Symposium (Simulation in Industry) 4-6 October 2007 Bergeggi (Sv), Italy |
|
Session CO-VII: Simulation & Optimization
AN OPEN-SOURCE DISCRETE EVENT SIMULATION-BASED OPTIMIZATION PACKAGE, Pasquale Legato, Roberto Trunfio
The need for an optimal management of manufacturing and other types of complex systems is well recognized. In this context, simulation is widely used for an accurate modeling of real systems, evaluating their performances and analyzing the effects of alternative conditions and organizational policies. General purpose and problem-oriented software packages have increased the diffusion of simulation as analysis tool in system management. Very interesting for real applications is a new methodology known as Simulation Optimization, in that it combines both simulation and optimization techniques with the objective of optimizing the performance measures of a system. Some optimization features have added in commercial simulation software, but they are not usually customizable. We describe the architecture of an optimization package based upon DEOS, an open-source discrete-event simulation platform. Finally, a case study on a manufacturing system is proposed to show the effectiveness of the simulation optimization package under development.
HELPDESK MODELING AND SIMULATION WITH DISCRETE EVENT SYSTEMS AND FUZZY LOGIC, I. Castilla, R. Muñoz, P. Baquero, R. M. Aguilar
Helpdesks are a fundamental component of the IT Infrastructure of any organization. A correct design and management of this service improve the performance of the whole organization. Managers are always looking for solutions which can support their decision making process. Discrete Event Simulation is becoming a widely spread tool which helps in the analysis and forecasting of the helpdesk behavior. However, the addition of some helpdesk processes, which are commonly expressed in uncertain or vague terms, could be unnecessarily complicated if only simulation is used. Fuzzy Logic can fill this hole left by simulation. This paper presents how Discrete Event Simulation and Fuzzy Logic can be used together to cope with the Quality of Service requirements of an organization.
HEURISTIC OPTIMIZATION METHODS FOR THE TUNING OF INPUT PARAMETERS OF SIMULATION MODELS, Michael Affenzeller, Gabriel Kronberger, Stephan Winkler, Mihaela Ionescu, Stefan Wagner
This paper proposes an approach combining Evolutionary Computation and Simulation: An extended problem representation as well as solution manipulation concepts for Evolution Strategies are proposed with the aim to fulfill the needs of input parameter optimization of simulation models; moreover, this approach is not restricted to real-valued parameters. The usage of a parallel infrastructure is also proposed and planned to be realized based on the HeuristicLab framework for heuristic optimization.
USING THE DESIGN OF EXPERIMENT FOR SIMULATION OF FACTOR'S BEHAVIOR-A CASE STUDY ON INORGANIC CHEMISTRY, Mahdi Bashiri, Sheida Ansar
Statistical designs of experiments refer to the process of planning the experiment so that appropriate data that can be analyzed by statistical methods will be collected, resulting in valid & objective conclusion. In this paper the design of experimental has been presented as a proper tool for simulation of factor's behavior in experiment. Finally a case study on inorganic chemistry with responses & factors has been presented. The result shows that the Solvent type (C) has a significant effect in this experiment and also Amin structure and Intermediate type have interaction with Solvent type.
VALUE-BASED SOFTWARE ENGINEERING USING THE ADAPTED TWIN-PEAKS MODEL, Herwig Mayr
We present our extension of the Twin Peaks Model that is suited as model for a slim, efficient management information tool for interactive-incremental (“agile”) software projects. Our Adapted Twin Peaks Model allows the depiction of product progress in relation to requirement changes, thus giving the project manager a quick inspection tool for analyzing project progress, focusing on the increase of customer value during the software engineering process (“value-based software engineering”). We give details on defining a suitable metrics for representing customer (i.e. business) value in relation to development effort and give examples, how the Adapted Twin Peaks Model can be used for determining project characteristics measures for validation & verification, economic feasibility, and customer satisfaction, among others. This is an ongoing project. Therefore, we give an outlook onto next investigations and features to be realized based upon the model.
|
