Abstract
Release planning is of key importance for incremental software product development. With the increasing size and complexity of software products, as well as with the growing demands for transparency and objectivity of decision-making, intuition alone is no longer sufficient for release planning. EVOLVE+ is a systematic method for release planning, which combines the strengths of formalization and computational efficiency with the expertise of the human experts. From a given set of candidate features, the most attractive ones are selected and assigned to a sequence of releases. In this paper, we consider three extensions to the current systematic planning approach EVOLVE+: (a) value functions describing the estimated value of the feature are continuous functions of time, (b) the actual release dates are no longer fixed but can be varied in some pre-defined interval. As a consequence, the available resource capacities are also functions of time. (c) Calculation of trade-off solutions balancing the risk with the potential additional value of early release. All three extensions substantially increase the complexity of the release planning problem. We have developed a solution method using genetic algorithms that is able to accommodate the additional complexity of the advanced model. A hypothetical case study is conducted as a proof-of-concept for the applicability of the method. The applicability of the method is demonstrated by analyzing six detailed planning scenarios.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call