A multi-attribute approach for release time and reliability trend analysis of a software

Abstract

In a Software Development Life Cycle the testing phase is given a lot of importance. But testing cannot be done indefinitely due to many reasons ranging from marketing considerations to increase in cost. Hence it is pertinent to find the optimal release time during testing phase. Firms routinely face the challenging decision of when to stop testing and release the product in the market. It has been observed that the pace of introduction of new software is much higher in recent years in comparison to any time in the past. Thus to get the competitive edge it is critical to know about the optimal entry time. Too late an entry is likely to lead to significant loss of opportunity. On the other hand early introduction of a product might hinder its growth due to lack of receptiveness of users towards new technology. Many release time problems with optimization criteria like cost minimization, reliability maximization and budgetary constraints etc. have been discussed in the literature. The release time problem discussed in this paper depends on the combined effect of cost and the Failure Intensity function; therefore we formulated an optimal release planning problem based on multi-attribute utility theory. A numerical illustration is provided towards the end of the paper. Furthermore, a trend line for the reliability of software has also been plotted to show the behaviour of reliability.