Automatic Algorithm Selection for Complex Simulation Problems
To select the most suitable simulation algorithm for a given task is often difficult. This is due to intricate interactions between model features, implementation details, and runtime environment, which may strongly affect the overall performance. An auto
- PDF / 5,853,553 Bytes
- 387 Pages / 419.528 x 595.276 pts Page_size
- 15 Downloads / 230 Views
VIEWEG+TEUBNER RESEARCH
Roland Ewald
Automatic Algorithm Selection for Complex Simulation Problems With a foreword by Prof. Dr. Adelinde M. Uhrmacher
VIEWEG+TEUBNER RESEARCH
Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available in the Internet at http://dnb.d-nb.de.
Dissertation Universität Rostock, 2010
1st Edition 2012 All rights reserved © Vieweg+Teubner Verlag | Springer Fachmedien Wiesbaden GmbH 2012 Editorial Office: Ute Wrasmann | Anita Wilke Vieweg+Teubner Verlag is a brand of Springer Fachmedien. Springer Fachmedien is part of Springer Science+Business Media. www.viewegteubner.de No part of this publication may be reproduced, stored in a retrieval system or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior written permission of the copyright holder. Registered and/or industrial names, trade names, trade descriptions etc. cited in this publication are part of the law for trade-mark protection and may not be used free in any form or by any means even if this is not specifically marked. Cover design: KünkelLopka Medienentwicklung, Heidelberg Printed on acid-free paper Printed in Germany ISBN 978-3-8348-1542-2
Foreword Simulation, an experiment performed with a model, belongs to the daily work of most scientists and practitioners in industry alike. Aimed at supporting the understanding, the analysis, and/or the design of complex dynamic systems, simulation belongs to the methodological toolbox of natural sciences, engineering, but also medicine, sociology, economy, and demography. The diversity of application areas and intentions of simulation studies is reflected in a plethora of available methods. “No silver bullet does exist” — this observation of Brooks, referring to software engineering in general, fits also well to simulation methods. Different models, infrastructures, and user preferences ask for different kinds of simulators. The performance of one method, e.g., in terms of execution speed, storage consumption, or accuracy of the results, might vary significantly from one situation to the next. Thus, users interested in performing simulation studies with their model are faced with the problem of how to select among existing methods the most suitable one, and developers of simulation methods are faced with the problem of how to evaluate the performance of their newly developed method in comparison to others. Those are daunting tasks, as most simulation methods are also highly configurable. However, solving these tasks is also important, as it will determine the quality of simulation studies and their results to a large degree. Roland Ewald’s book on simulation algorithm selection contributes to this quest. It shows how methods from machine learning, portfolio theory, experiment design, adaptive software, and simulation algorithms can be combined to develop new approaches for
Data Loading...
