Project Scope Partitioning by Clustering Features into Releases of Long R&D Projects

R&D projects are characterized by a long planning horizon, which entails the policy of release management. The intermediate releases enable the organization to maximize the value for a given investment. Myopic version of this problem is known as the Next Release Problem (NRP). A central issue addressed by these projects is determining which features should be included in the next release. The choice of features impacts the value of the release, but also impacts the required workload, and future development of other features. NRP can be expanded to include the later releases. This problem is NP-hard and thus cannot be solved analytically. In this work we apply a simple clustering algorithm, based on novel similarity coefficients to reduce complexity. Our goal is to provide a near-optimal yet simple method for quantitatively determining the feature content of all project releases.