Multinomial-exponential reliability function: a software reliability model |
| |
Authors: | Amalio Saiz de Bustamante Barbara Saiz de Bustamante |
| |
Affiliation: | a Department of Nuclear Engineering, Universidad Politecnica de Madrid, 28040, Madrid, Spain;b Department of Quality, UN12,100 Ortega y Gasset, 28006, Madrid, Spain |
| |
Abstract: | The multinomial-exponential reliability function (MERF) was developed during a detailed study of the software failure/correction processes. Later on MERF was approximated by a much simpler exponential reliability function (EARF), which keeps most of MERF mathematical properties, so the two functions together makes up a single reliability model. The reliability model MERF/EARF considers the software failure process as a non-homogeneous Poisson process (NHPP), and the repair (correction) process, a multinomial distribution. The model supposes that both processes are statistically independent.The paper discusses the model's theoretical basis, its mathematical properties and its application to software reliability. Nevertheless it is foreseen model applications to inspection and maintenance of physical systems. The paper includes a complete numerical example of the model application to a software reliability analysis. |
| |
Keywords: | Non-homogeneous Poisson process Exponential reliability function Software reliability model |
本文献已被 ScienceDirect 等数据库收录! |
|