Electronic Thesis and Dissertation Repository

Degree

Master of Engineering Science

Program

Electrical and Computer Engineering

Supervisor

Luiz Fernando Capretz, PhD, P.Eng

Abstract

Software development can be considered to be the most uncertain project when compared to other types of projects due to the uncertainty in the customer requirements, the complexity of the process, and the intangible nature of the product. In order to increase the likelihood of success in managing a software project, the project manager(s) must invest more time and effort in the project planning phase. The two main activities in project planning phase are effort estimation and risk management which has to be executed together because the accuracy of the effort estimation is highly dependent on the size and nature of the project risks.

However, as is common practice in a software development project, effort estimation and risk management are often disconnected from each other and most of the software effort estimation methodologies, which include the COCOMO model, provide a fixed rather than an approximate value, and consequently the existing effort estimation approach has failed to provide a reliable reference for project manager due to its lack of accuracy.

This thesis introduces the Fuzzy Expert-COCOMO Model, the Risk Assessment and Effort Contingency Model based on cost factors and fuzzy technique, which has the capability to not only integrate the effort estimation and risk assessment activities into the initial project planning phase but also to provide the essential information about the estimated effort, the project risks, and the effort contingency allowance required to accommodate the identified risk.

A validation of this model using project data set shows that the new approach is capable of improving the sensitivity of project risk assessment using Expert-COCOMO methodology and also provides a higher level of effort prediction performance compare to the existing COCOMO-II effort estimation.

Share

COinS