| This research focuses on the issues of risk arising from multiple projects undertaken by a single company. Projects are dependent on a number of risk factors that interact in complex ways and that make the outcome of a project uncertain. In these cases, it may be possible for contractors to manage these risks by choosing corporate-level management strategies. By bearing these considerations in mind, this thesis develops a risk analysis methodology to serve as a support for managers analyzing a project's risks; responses can be for the specific project only or across many projects.;The methodology in this thesis is built on the work breakdown structure (WBS) of a project, in which a project is defined as a portfolio of its component work packages. A single-index, linear model is then developed to estimate the uncertainty of a work package, which depends on the basic concept of the market model in financial portfolio theory. The key parameter of the developed model is beta, which is a regression coefficient relating the performance of a work package to the overall performance of completed projects. The beta estimates the co-variance among work packages, or a portion of project risk that arises from common risk factors to projects. Moreover, it is used to predict the change in profile of a project's uncertainty when risk-control is evaluated; estimating the value of controlling a risk is a principal contribution of this research.;This methodology is uniquely formulated in that the risk is estimated only with the components in WBS and the associated cost information, instead of relying on external, explanatory variables. Such a model, built with rigorous statistical inference, provides a systematic and reliable approach for project risk analysis. In practice, it not only facilitates the knowledge acquisition process, but it also considerably reduces the cost of information. Thus, the application of this methodology will be of great valuable to both industry and the academic community. |
