A comparative study of the relative performance and real-world suitability of optimization approaches for Human Resource Allocation

Al Khatib, Sultan (2018) A comparative study of the relative performance and real-world suitability of optimization approaches for Human Resource Allocation. Doctoral thesis, University of East Anglia.

[thumbnail of SultanMAALKhatib-AComparativeStudyOfRelativePerformanceAndReal-WorlSuitabilityOfHRA-PhD-Thesis-09-18.pdf]
Preview
PDF
Download (3MB) | Preview

Abstract

The problem of Staffing and Scheduling a Software Project (SSSP), where we consider Human Resource Allocation (HRA) to minimize project time, offers a management challenge for Project Managers (PM’s). Unlike the general HRA problem, SSSP involves determination of the assignment of a fixed amount of resources to teams and the allocation of these teams to project’s jobs. SSSP problem arises across a diverse range of resources’ and project characteristics (discrete variables), and this variety has offered a wide range of HRA methods.

The general consensus is that the benchmark for SSSP are Meta-heuristic optimization techniques using deterministic or stochastic simulation of time. However, different HRA methods and project attributes are considered by SSSP approaches, and their solutions need to be compared against each other. The majority of SSSP approaches provide their approximation using Genetic Algorithm (GA) validated by a synthetic data or empirical method such as Quasi-experiment. Limited studies offer the comparison between these SSSP approaches, either by a comprehensive survey or systematic literature review for qualitative concepts.

We aim to answer a set of research questions including: what is the best way to show the quality and performance differences between SSSP approaches? And, are these SSSP approaches suitable for industrial adoption? Our thesis is that the best methodology is to identify according to the conceptual models used by the approaches a set of challenging data levels. In support of our thesis, we propose a systematic benchmarking and evaluation approach that encompass the data levels, and a set of quality measures. Next, we propose an empirical study that assess how PMs from software industry perform the allocation given the same datasets. The results of both works demonstrate significant differences between the approaches, highlighted four methods that advances the research filed, and provide interesting discussion on the PMs’ practices on SSSP.

Item Type: Thesis (Doctoral)
Faculty \ School: Faculty of Science > School of Computing Sciences
Depositing User: Users 9280 not found.
Date Deposited: 10 Sep 2018 12:51
Last Modified: 10 Sep 2018 12:51
URI: https://ueaeprints.uea.ac.uk/id/eprint/68211
DOI:

Actions (login required)

View Item View Item