Use this URL to cite or link to this record in EThOS: https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.807954
Title: A Bayesian approach for software release planning under uncertainty
Author: Oni, Olawole Stephen
Awarding Body: UCL (University College London)
Current Institution: University College London (University of London)
Date of Award: 2020
Availability of Full Text:
Access from EThOS:
Full text unavailable from EThOS. Please try the link below.
Access from Institution:
Abstract:
Release planning — deciding what features to implement in upcoming releases of a software system— is a critical activity in iterative software development. Many release planning methods exist but most ignore the inevitable uncertainty of future development effort and business value. The thesis investigates how to analyse uncertainty during release planning and whether analysing uncertainty leads to better decisions than if uncertainty is ignored. The thesis’s first contribution is a novel release planning method designed to analyse uncertainty in the context of the Incremental Funding Method, an incremental cost-value based approach to software development. Our method uses triangular distributions, Monte-Carlo simulation and multi-objective optimisation to shortlist release plans that maximise expected net present value and minimise investment cost and risk. The second contribution is a new release planning method, called BEARS, designed to analyse uncertainty in the context of fixed-date release processes. Fixed-date release processes are more common in industry than fixed-scope release processes. BEARS models uncertainty about feature development time and economic value using lognormal distributions. It then uses Monte-Carlo simulation and search-based multi-objective optimisation to shortlist release plans that maximise expected net present value and expected punctuality. The method helps release planners explore possible tradeoffs between these two objectives. The thesis’ third contribution is an experiment to study whether analysing uncertainty using BEARS leads to shortlisting better release plans than if uncertainty is ignored, or if uncertainty is analysed assuming fixed-scope releases. The experiment compares 5 different release planning models on 32 release planning problems. The results show that analysing uncertainty using BEARS leads to shortlisting release plans with higher expected net present value and higher expected punctuality than methods that ignore uncertainty or that assume fixed-scope releases. Our experiment therefore shows that analysing uncertainty can lead to better release planning decisions than if uncertainty is ignored.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.807954  DOI: Not available
Share: