Generalised multi release framework for fault determination with fault reduction factor

The huge dependability on software systems has led to the need of reliable software in a short span of time. One of the ways to achieve this is to provide a series of versions of the software system. Thus, firms release the first software version with the desired level of reliability in which all faults are not removed. The remaining faults are then removed during the software up-gradation. The main focus in this paper is to deal with multi-release modelling incorporating the concept of fault reduction factor (FRF). FRF is a significant factor in determining the software reliability and is defined as the net reduction in the number of faults in proportion to the total number of experienced failures. In this paper, multi-release models are proposed incorporating a constant FRF with imperfect debugging, change point and effort function. The parameter estimation is done on the four releases of Tandem dataset. Numerical illustrations are given to validate the proposed model.