The problem of assigning most appropriate requirements to a series of releases of a software system is difficult to solve due to uncertainty from several sources, for example, the preferences of different stakeholders. We present a solution to this problem by providing a flexible release planning procedure using a solution generation engine, ReleasePlanner®, and so-called e-assistants. In our iterated e-release planning process, e-assistants present to their human stakeholders solutions to variants of the problem instance. By selecting the best suited solutions, the stakeholders allow the e-assistants to elicit more and more their implicit preferences. To guarantee termination, in each round the assignment of some requirements to releases is fixed, based on analysing concordance and non-discordance of assignments between the preferred candidate solutions.
Generated: 14/1/2007