GRL goal analysis using zero-sum game theory

Abstract

In Requirements Engineering (RE), goal models are used to represent stakeholder objectives, which are also known as system requirements or system goals. Stakeholder requirements are of two types: functional requirements and non-functional requirements. Goal models are analysed to find out suitable functional requirements amongst ensemble of all functional requirements. RE literature has addressed both qualitative and quantitative methods for performing goal analysis. Recently, operation research techniques have been used for performing optimal goal analysis. The existing optimisation approaches focus on maximising objective functions. But real world problems involve optimisation of both maximising and minimising objective functions simultaneously. In this study, a game theory based approach is used for solving simultaneous optimisation of both maximum and minimum objective functions in goal models. The proposed approach is applied to Goal-Oriented Requirement Language (GRL) framework which is perceived as a standard for goal-modelling. The practicality of the proposed approach was assessed by running the case studies in a simulated environment using JAVA Eclipse combined with IBM Cplex tool. The results showed that the proposed approach aids in the analysis of goals in goal models with opposing objective functions.