A Grid-Inspired Mechanism for Coarse-Grained Experiment Execution

Leye, Stefan; Himmelspach, Jan; Jeschke, Matthias; Ewald, Roland; Uhrmacher, Adelinde M.

doi:10.1109/ds-rt.2008.33

Cited by 18 publications

(17 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The development of an experimentation layer for JAMES II has been a group effort right from the beginning [68,74,132,201]. It is a complex and feature-rich part of JAMES II that cannot (and need not) be described here in full detail; the following discussion is restricted to a brief overview of the main concepts and then centers on those entities that are exploited or extended later on.…”

Section: The Experimentation Layer Of James IImentioning

confidence: 99%

“…The execution as such is transparent to the experiment that generates the simulation configurations, which makes it easy to support different execution strategies for different kinds of infrastructure, as discussed in [201]. Again, the flexibility of choosing different simulation runners is realized by defining a corresponding plug-in type and plug-ins.…”

Section: Experiments Controlmentioning

confidence: 99%

“…• Asynchrony: Simulation runners may exploit multiple resources and schedule simulation executions to run concurrently, i.e., they may conduct parallel independent replications [201]. Hence, a policy may have to decide which arm to pick before it received all -or, in fact, any -reward, as illustrated in figure 7.…”

Section: 1 P 4)mentioning

confidence: 99%

“…The adaptive simulation runner reuses the functionality of the parallel simulation runner presented in [201]. A new plug-in type for multi-armed bandit policies that minimize reward is introduced, it subsumes implementations of the interface IMinBanditPolicy.…”

Section: General Structurementioning

confidence: 99%

See 3 more Smart Citations

Automatic Algorithm Selection for Complex Simulation Problems

Ewald¹

2012

Self Cite

View full text Add to dashboard Cite

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 automated selection of simulation algorithms supports users in setting up simulation experiments, without demanding expert knowledge on simulation.The first part of the thesis surveys existing approaches to solve the algorithm selection problem and discusses techniques to analyze simulation algorithm performance. A unified categorization of existing algorithm selection techniques is worked out, as these stem from various research domains (e.g., finance, artificial intelligence).The second part introduces a software framework for automatic simulation algorithm selection and describes its constituents, as well as their integration into the modeling and simulation framework JAMES II. The implemented selection mechanisms are able to cope with three situations: a) no prior knowledge is available, b) the impact of problem features on performance is unknown, and c) a relationship between problem features and algorithm performance can be established empirically.An experimental evaluation of the developed methods concludes the thesis. It is shown that an automated algorithm selection may significantly increase the overall performance of a simulation system. Some of the presented mechanisms also support the research on simulation methods, as they facilitate their development and evaluation. ZusammenfassungDie Auswahl des passendsten Simulationsalgorithmus für eine bestimmte Aufgabe ist oftmals schwierig. Dies liegt an der komplexen Interaktion zwischen Modelleigenschaften, Implementierungsdetails und der Laufzeitumgebung, welche die Gesamtleistung stark beeinflussen kann. Eine automatisierte Auswahl von Simulationsalgorithmen unterstützt die Anwender eines Simulationssystems, indem sie eine geeignete Konfiguration des Systems ohne zusätzliches Fachwissen aus dem Bereich Simulation ermöglicht.Der erste Teil der Arbeit befasst sich eingehend mit Vorarbeiten zur automatischen Algorithmenauswahl, sowie mit der Leistungsanalyse von Simulationsalgorithmen. Eine einheitliche Kategorisierung wird erarbeitet um die aus den verschiedensten Fachgebieten (z.B. Finanzwesen, Künstliche Intelligenz) stammenden Ansätze zur Algorithmenauswahl einzuordnen.Der zweite Teil der Arbeit stellt ein Rahmenwerk zur automatischen Auswahl von Simulationsalgorithmen vor und beschreibt dessen Bestandteile, einschließlich ihrer Integration ins Modellierungsund Simulationsrahmenwerk JAMES II. Die realisierten Auswahlmechanismen beziehen sich dabei auf drei Situationen: a) keinerlei Vorwissen ist vorhanden, b) der Einfluss von Problemeigenschaften auf die Leistung ist unbekannt, und c) ein Zusammenhang zwischen Problemeigenschaften und Algorithmenleistung konnte empirisch ermittelt werden.Eine experimentelle Untersuchung der entwickelten Methoden schließt die Arbeit ab. Es wird gezeigt, dass eine aut...

show abstract

Section: The Experimentation Layer Of James IImentioning

confidence: 99%

Section: Experiments Controlmentioning

confidence: 99%

Section: 1 P 4)mentioning

confidence: 99%

Section: General Structurementioning

confidence: 99%

See 2 more Smart Citations

Automatic Algorithm Selection for Complex Simulation Problems

Ewald¹

2012

Self Cite

View full text Add to dashboard Cite

show abstract

“…Em Leye et al (2008) o autor descreve uma abordagem para execução de simulações estocásticas em ambiente de computação em grade. O conceito principalé a criação de um servidor de simulação, o qualé instanciado no recurso capaz de processar e outros serviços que servem como estágios para a divisão e gerenciamento dos experimentos a serem executados.…”

Section: James IIunclassified