Software reliability modelling and evaluation via Markov and semi-Markov processes

Abstract

This thesis develops Markov and semi-Markov models for quantitative evaluation of software reliability. Two measures software reliability termed as operational reliability and operational availability have been introduced which provide better figures of merit than reliability. Quantitative evaluation of total expected cost and the cost due to failures in control transfer of a computer program are described and the overall failure process of the program is developed. Program examples are presented to illustrate the procedure presented in this thesis. Some possible future extensions of the work are also examined.