Towards Multi-objective Optimisation of Quantitative Goal Models using Constraint Programming

Goal Model are widely used to capture system goals and refine them into operational requirements assigned to human, hardware or software. Such models support alternative goal refinements resulting in a potentially large design space to explore. A given design can be quantitatively evaluated in terms of its fulfilment of a set of non-functional requirements (e.g. reliability, performance) or business goals (e.g. costs, stakeholder satisfaction). Optimisation techniques can be used to explore the design space to determine an optimal design according to a single objective like the cost but also according to multi-objective techniques to propose a set of Pareto-optimal solutions in which a best solution can be selected. In this paper, we show how to translate a goal-oriented requirements model, expressed in the KAOS notation, into a constraint programming (CP) problem. The OscaR.CP engine is used to get, from all alternatives co explored, either global or Pareto-optimal solutions. Our method is implemented as a tool plugin of a requirements engineering platform and is benchmarked on the classical meeting scheduler case study.