Modelling multi up gradations and related release time problems

Abstract

Testing forms an important aspect of software development life cycle and plays an imperative role in defining the quality of a software. But, due to limited resources and time constraint, firms have to decide upon their upper limits while investing. Moreover, due to fierce competition in the market, software nowadays are short-lived. Hence, instead of delivering the complete product in a single development cycle, firms plan periodic releases of software and come up with multiple successive up gradations to survive. But at the same time newly added features increase the software complexity which further magnifies faults in the software. To follow the effect of faults generated in the software due to functional enrichment, a multi up gradation software reliability model has been discussed in this paper. This model identifies left over faults from the operational phase of previous version of the software, which are debugged during the testing phase of the next version. Multi up gradation elucidate the real world approach for software development. But, it accompanies other concerns as well. One such matter of contention is the optimal release time problem. Many models have been designed in the past which talks about when to stop testing and release the software to the users. In this paper we have discussed a unified approach to address this crucial argument of when to stop testing multi up-gradation of software under the assumption that the software is supported till the warranty period is over. While discussing the cost model for each up-gradation it was considered that some of the remaining faults of previous release are reported and removed partly during the testing period and partly during the warranty period of new up-gradation. Further, we discussed about estimating optimal software release time using Multi Attribute Utility Theory. More precisely, three considerable attributes, viz. Reliability, Cost and Detection rate indicator are employed in order to determine the optimal release time. Finally, we have examined the multi attribute utility theory approach to find optimal release time in case of multi-release of a software with constraints on reliability & cost. The parameters of the model are estimated using Statistical Package for Social Sciences (SPSS) based on real life data set & optimum stopping time for each version of software is obtained using Maple software.