Multi-objective dynamic software project scheduling: an evolutionary approach for uncertain environments

Nigar, Natasha (2021). Multi-objective dynamic software project scheduling: an evolutionary approach for uncertain environments. University of Birmingham. Ph.D.

Preview

Nigar2021PhD.pdf
Text - Accepted Version
Available under License All rights reserved.
Download (2MB) | Preview

Abstract

Software project scheduling (SPS) in an uncertain and dynamic environment is critical for within budget and on time completion of real-world projects. This is an important issue in the software engineering practices, as budget and effort must be managed well enough for successful project completion. This deals with suitable allocation of employees to tasks in a software project with an aim to minimize project cost and schedule. Most of the existing research address this problem by considering only static and deterministic scenarios with no dynamic disruptions. However, the SPS has emerged as a dynamic scheduling problem which is challenged by inherent agility for medium to large scale projects. It requires not only the capability to handle dynamic events but also multiple objectives to ensure successful project completion. The increasing trend of cloud based software as a service (SAAS) solutions (large-scale complex projects) also requires projects on schedule and within budget.

In this thesis, we address this challenge (handling of dynamic events potentially capable to disrupt project quality, schedule, and cost) by formulating the software project scheduling problem as an optimization problem that regenerates a feasible software project schedule. We present three multi-objective (project duration, cost, robustness, and stability) models, subjected to variety of practical constraints, to deal with dynamic events. First, we present a model to deal with ‘Employee Turnover’ dynamic event that handles both employee rescheduling and recruitment scenarios, along with a multi-objective evolutionary algorithm. Second, we tackle another dynamic event ‘new employee addition’ by proposing a multi-objective evolutionary algorithm as novel heuristic approach. Third, we propose a new model for the SPS problem by considering an important human factor i.e. ‘employee experience’.

The work in this thesis presents our first attempts to address some of the most challenging problems in software project scheduling under dynamic environment. The proposed evolutionary algorithmic approaches and models have been evaluated against existing state-of-the-art algorithms.

Type of Work:

Thesis (Doctorates > Ph.D.)

Award Type:

Doctorates > Ph.D.

Supervisor(s):