Risk-Aware Multi-stakeholder Next Release Planning Using Multi-objective Optimization

[Context and motivation]: Software requirements selection is an essential task in the software development process. It consists of finding the best requirement set for each software release, considering several requirements characteristics, such as precedences and multiple conflicting objectives, such as stakeholders’ perceived value, cost and risk. [Question/Problem]: However, in this scenario, important information about the variability involved in the requirements values estimation are discarded and might expose the company to a risk when selecting a solution. [Principal ideas/results]: We propose a novel approach to the risk-aware multi-objective next release problem and implemented our approach by means of a satisfiability modulo theory solver. We aim at improving the decision quality by reducing the risk associated with the stakeholder dissatisfaction as related to the variability of the value estimation made by these stakeholders. [Contribution]: Results show that Pareto-optimal solutions exist where a major risk reduction can be achieved at the price of a minor penalty in the value-cost trade-off.