Planning optimal agile releases via requirements optimization

This paper focuses on improving requirements quality in agile projects by determining requirements prioritization. Current methods suggest to take into account business value in order to determine the requirements priority rank. In practice it was observed that many other factors enter into the equation, such as implementation cost and functionality dependencies. Since agile methods suggest that priority should be customer/user's prerogative, taking all relevant factors becomes challenging without decision supporting tools. Our research question is the following: How can we formulate the agile release decision problem, and which computations can we do over requirements models to recommend solutions to that decision problem? Our contributions are the following: (i) we formulate this agile release decision problem as an optimization problem, (ii) we provide a modelling language to represent instances of this problem as requirements models, and (iii) we describe an online tool to make the models and solve the resulting optimization problem instances.