by Fedon Kadifeli, B.S. in CMPE., Boğaziçi University, 1987
Submitted to the Institute for Graduate Studies in Science and Engineering in partial fulfillment of the requirements for the degree of Master of Science in Computer Engineering, Boğaziçi University, 1989
APPROVED BY
Doç. Dr. Oğuz Tosun
(Thesis Supervisor)
Doç. Dr. Selahattin Kuru
Doç. Dr. Süleyman Özekici
DATE OF APPROVAL
12/09/1989
YÖK Record Number
8007
I would like to thank Dr. Ufuk Çağlayan who proposed the thesis topic and advised me in all stages of this work. Thanks are also due to Dr. Oğuz Tosun, Neşe Akkaya, and Ayşe Bayazıtoğlu who supplied some of the references used in this thesis. I thank also Dr. Oğuz Tosun, Dr. Selahattin Kuru, and Dr. Süleyman Özekici for their careful reviews.
In this thesis a model that can be used for conventional software reliability analysis and prediction is described. The proposed model has been applied on a large-scale distributed commercial software system which is already operational. The model is characterized by the gamma distribution having three unknown parameters. The parameters are estimated from failure time data by using the method of maximum likelihood estimation. The model is shown to fit the data better than other well-known models in the absence of additional information on the software system being analyzed.
Bu tezde geleneksel yazılım güvenilirliği çözümlemesi ve önkestiriminde kullanılabilecek bir model anlatılmaktadır. Önerilen model işletimsel durumda olan büyük ölçekli dağıtımlı ticari bir yazılım sistemi üzerinde uygulanmıştır. Model üç bilinmeyen parametresi olan gama dağılımı ile tanımlanmaktadır. Bu parametreler başarısızlık zaman verilerinden en çok olabilirlik kestirimi yöntemi kullanılarak kestirilmektedir. Modelin, çözümlenmekte olan yazılım sistemi üzerine daha fazla bilgi yokluğunda, diğer bilinen modellerden verilere daha iyi uyduğu gösterilmektedir.
ACKNOWLEDGEMENTS
ABSTRACT
ÖZET
LIST OF FIGURES
LIST OF TABLES
LIST OF SYMBOLS
I. INTRODUCTION
II. SOFTWARE QUALITY AND RELIABILITY
III. SOFTWARE RELIABILITY MODELING
IV. PARAMETER ESTIMATION
V. DESCRIPTION OF COLLECTED DATA
VI. RELIABILITY MODEL
VII. CONCLUSIONS
A. OPTIMIZATION ALGORITHM
B. INDEX
BIBLIOGRAPHY
REFERENCES NOT CITED
The full text of the thesis (in Adobe Acrobat Reader format) can be extracted from the archive file msthesis.zip (277 KiB). If you encounter any difficulties in downloading or reading the document, please contact me.