Optimal software release time determination with risk constraint

Abstract

For a software development project, when to stop testing the software and release it for operation is of great importance as it impacts both the software reliability and the total cost of the project. Software release time determination, therefore, has attracted a lot of research in the past two decades and several new cost models have been developed in the literature recently. In most research on this topic, the approach taken is to minimize the expected total cost (ETC) of the software project, or further consider the software reliability requirement. However, because the actual total cost (ATC) of the software project is a random variable, minimization of the ETC does not guarantee that the ATC will be near this minimum. In fact, there exists certain risk that the ATC may exceed the ETC to an intolerable extent, which, despite its importance, has not been addressed in most related research. In this paper, we study the above mentioned risk problem for software release time determination and propose a new approach which could be helpful for management to control the risk of the project being over-budget. A numerical example is given to illustrate the solution procedures of the proposed approach.