Software reliability and cost models with environmental factors

This dissertation aims to advance the theories and methods of software reliability and cost assessment in a systematic and realistic way. The stochastic behaviors of software development process are extensively investigated and the environmental factors that profile the development process are thoroughly studied. Software cost models integrating warranty cost and risk cost are proposed.; Today software plays such a critical role in both safety-critical and civilian systems that our modern society relies heavily on its successful performance. Therefore, from both the users' and developers' point of view, there is a great need for high quality software. Software development is a complex process, in which many environmental factors are involved. The environmental factors include characteristics of the software itself (e.g., complexity); the development environment (e.g., people and tools); and all other factors influencing the whole software development process. These factors have significant impacts on the final quality of the software product and need to be taken into consideration. In this dissertation, thirty-two environmental factors have been defined and their impacts on software reliability have been studied. Software reliability models incorporating these factors using the non-homogeneous Poisson process are proposed to enhance the accuracy of software reliability assessment.; Software cost models in which several issues such as error removal cost during testing and operational phases, warranty cost, and the risk cost due to software failures are contemplated. The optimal software release policies that can be used to determine when to stop testing the software are proposed. The proposed models have been validated using software failure data from real software applications.