A Software Reliability Growth Model for Vital Quality Metrics

A Non-Homogenous Poisson Process (NHPP) model whose failure intensity function has the same mathematical form as that of a generalized exponential function was proposed for application as a Software Reliability Growth Model (SRGM). However, in order to facilitate collecting quality metrics pertaining to the degree of imperfect or efficient debugging phenomena and the number of faults left in the software, in this paper the authors propose an extension to the above SRGM. This SRGM enables adequate goodness of fit statistic and predictive validity, even when the software projects witness learning phenomenon of the testing team, either imperfect or perfect or efficient software debugging phenomenon, as well as wide fluctuations in time between failures – either occurring alone or in combinations thereof. OPSOMMING 'n Nie-homogene Poissonproses (NHPP) waarvan die mislukkingsdigtheidsfunksie soortgelyk is aan 'n algemene eksponensiële funksie word voorgehou as 'n programmatuur-betroubaarheidsgroeimodel (PBGM). Die model lewer toereikende passingsgoedheid en voorspellingsgeldigheid onder uiteenlopende leereienskappe van toetsers, swak of goeie ontfouting van programmatuur, en groot verskille tussen waardes van tyd tussen mislukkings. Die outeurs stel ook voor dat die goedheid van ontfoutingsaksies gemeet word met behulp van 'n uitbreiding van die PBGM-model