Multi-Stakeholder Tensioning Analysis in Requirements Optimisation Yuanyuan Zhang CREST Centre University College London

Agenda Background Problem Solution Empirical Study Conclusion

Requirements Engineering Process Acquisition Modelling & Analysis Specification Validation & Verification Evolution & Management University College London Modelling& Analysis Evolution & Management Background ProblemSolutionEmpirical StudyConclusion

Requirements Selection & Optimisation Task Using prioritisation, visualisation, and optimisation techniques helps decision maker to select the optimal or near optimal subset from all possible requirements to be implemented. Background ProblemSolutionEmpirical StudyConclusion University College London

Goals Cost Revenue Stakeholder A Stakeholder B R5R4R3R2R1 Background Problem SolutionEmpirical StudyConclusion University College London

Search based Requirements Optimisation Use of meta-heuristic algorithms to automate and optimise requirements analysis process – Choose appropriate representation of problem – Define problem specific fitness function (to evaluate potential solutions) – Use search based techniques to lead the search towards optimal points in the solution space BackgroundProblem Solution Empirical StudyConclusion University College London

Representation Stakeholder: Weight: Requirements: Cost: BackgroundProblem Solution Empirical StudyConclusion University College London

Representation Each stakeholder assigns a value to requirements : Each stakeholder has a subset of requirements that expect to be fulfilled denoted by The overall score of a given requirement can be calculated by: BackgroundProblem Solution Empirical StudyConclusion University College London

Data Set Generation & Initialisation Real World Data Sets 27 Combination Random Data Sets Other Random Data Sets Requirement.Number Requirement.Value Requirement.Cost Requirement.Dependency Stakeholder.Number Stakeholder.Weight Stakeholder.Subset Requirement-Stakeholder Matrix.Density Random.Distribution Requirement.Number Requirement.Value Requirement.Cost Requirement.Dependency Stakeholder.Number Stakeholder.Weight Stakeholder.Subset Requirement-Stakeholder Matrix.Density Random.Distribution initialise input generate Matlab Files.mat.dat Matlab Files.mat.dat format process BackgroundProblem Solution Empirical StudyConclusion University College London

Interaction Management Multi-Stakeholder Analysis Value/Cost Trade-off Analysis Basic Value/Cost Today/Future Importance Fitness-Invariant Dependency Fitness-Affecting Dependency Tensioning Analysis Fairness Analysis Iterate ? Chang e? Yes No Data Sets Regeneration Search Algorithms NSGA-II Archive-based NSGA-II Two-Achieve Pareto GA Single Objective GA Greedy* Random* Statistic Analysis * Strictly speaking, these are not search algorithms. Spearman’s Rank Correlation ANOVA Analysis Requirements Selection Process

Result Representation and Visualisation Results Requirements Subsets for Release Planning Insight Characteristic of Data Sets Performance of the Algorithms represent & visualise 2D and 3D Pareto Fronts Kiviat Diagrams Convergence Marked Line Charts Diversity communicate & feedback King’s College London BackgroundProblem Solution Empirical StudyConclusion

Visualisation Pareto Optimal Front BackgroundProblem Solution Empirical StudyConclusion University College London

The problem is to select a set of requirements that maximise the total value to each stakeholder, which is expressed as a percentage. The model of fitness functions represented as: Maximise subject to University College London Fitness Functions BackgroundProblem Solution Empirical StudyConclusion

Data Sets Used 2. Greer and Ruhe Data Set: 20 Requirements and 5 Stakeholders 1. Motorola Data Set: 35 Requirements and 4 Stakeholders King’s College London BackgroundProblemSolution Empirical Study Conclusion

Data Sets Used Combination Levels of Random Data Sets: the No. of requirements the No. of stakeholders the density of the stakeholder-requirement matrix King’s College London BackgroundProblemSolution Empirical Study Conclusion

Multi-Stakeholder Kiviat Diagram BackgroundProblemSolution Empirical Study Conclusion 0% 75% 100% 50% 25% Sta. 1 Sta. 4 Sta. 3 Sta. 2 University College London

Multi-Stakeholder Tensioning Analysis Motorola data set 30% Budgetary Resource Constraint 70% BackgroundProblemSolution Empirical Study Conclusion University College London

Solutions on the Pareto Front

Average Solutions

Tensions between the Stakeholders’ Satisfaction for Different Budgetary Resource Constraints

Multi-Stakeholder Tensioning Analysis Greer and Ruhe data set 30% Budgetary Resource Constraint 70% BackgroundProblemSolution Empirical Study Conclusion University College London

Algorithms’ Performance BackgroundProblemSolution Empirical Study Conclusion University College London

Algorithms’ Performance 1. The diversity of the Two-archive algorithm is significant in most cases 2. The Two-archive and NSGA-II algorithms always have a better convergence than the Random Search 3. The Two-Archive algorithm outperforms NSGA-II and Random Search in terms of convergence in some case BackgroundProblemSolution Empirical Study Conclusion University College London

Conclusion Present the concept of multi-stakeholder tensioning in requirements analysis Treat each stakeholder as a separate objective to maximise the requirement satisfaction Present the empirical study and statistical analysis Provide a simple visual method to show the tensioning University College London