Endre søk
RefereraExporteraLink to record
Permanent link

Direct link
Referera
Referensformat
  • apa
  • harvard1
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Annet format
Fler format
Språk
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Annet språk
Fler språk
Utmatningsformat
  • html
  • text
  • asciidoc
  • rtf
ON EFFICIENT AUTOMATED METHODS FOR SIMULATION OUTPUT DATA ANALYSIS
Blekinge Tekniska Högskola, Fakulteten för datavetenskaper, Institutionen för kommunikationssystem.
2014 (engelsk)Independent thesis Advanced level (degree of Master (Two Years))Oppgave
Abstract [en]

With the increase in computing power and software engineering in the past years computer based stochastic discrete-event simulations have become very commonly used tool to evaluate performance of various, complex stochastic systems (such as telecommunication networks). It is used if analytical methods are too complex to solve, or cannot be used at all. Stochastic simulation has also become a tool, which is often used instead of experimentation in order to save money and time by the researchers. In this thesis, we focus on the statistical correctness of the final estimated results in the context of steady-state simulations performed for the mean analysis of performance measures of stable stochastic processes. Due to various approximations the final experimental coverage can differ greatly from the assumed theoretical level, where the final confidence intervals cover the theoretical mean at much lower frequency than it was expected by the preset theoretical confidence level. We present the results of coverage analysis for the methods of dynamic partially-overlapping batch means, spectral analysis and mean squared error optimal dynamic partially-overlapping batch means. The results show that the variants of dynamic partially-overlapping batch means, that we propose as their modification under Akaroa2, perform acceptably well for the queueing processes, but perform very badly for auto-regressive process. We compare the results of modified mean squared error optimal dynamic partially-overlapping batch means method to the spectral analysis and show that the methods perform equally well.

sted, utgiver, år, opplag, sider
2014. , s. 148
Emneord [en]
Akaroa2, batch means, simulation output analysis, sequential coverage analysis, spectral analysis
HSV kategori
Identifikatorer
URN: urn:nbn:se:bth-5857Lokal ID: oai:bth.se:arkivex35B37F63EF222097C1257D3B004919CCOAI: oai:DiVA.org:bth-5857DiVA, id: diva2:833263
Uppsök
Technology
Veileder
Merknad
+420 723 771 283Tilgjengelig fra: 2015-04-22 Laget: 2014-08-21 Sist oppdatert: 2018-01-11bibliografisk kontrollert

Open Access i DiVA

fulltekst(8828 kB)398 nedlastinger
Filinformasjon
Fil FULLTEXT01.pdfFilstørrelse 8828 kBChecksum SHA-512
f4e2f54acf223c2f9409a8d390e995592523e63e146ab4bace5c53dffb4af317bbaa6d57e11313a766764697721c14f2a82b905d37b6b7958d896aec54065150
Type fulltextMimetype application/pdf

Av organisasjonen

Søk utenfor DiVA

GoogleGoogle Scholar
Totalt: 398 nedlastinger
Antall nedlastinger er summen av alle nedlastinger av alle fulltekster. Det kan for eksempel være tidligere versjoner som er ikke lenger tilgjengelige

urn-nbn

Altmetric

urn-nbn
Totalt: 319 treff
RefereraExporteraLink to record
Permanent link

Direct link
Referera
Referensformat
  • apa
  • harvard1
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Annet format
Fler format
Språk
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Annet språk
Fler språk
Utmatningsformat
  • html
  • text
  • asciidoc
  • rtf