Uncertainty In Cost Analysis

Generally, risk is considered to be a concept which is difficult to measure. Uncertainty in cost and risk has always been a complex part of every project. The purpose of cost analysis is to help decision makers to recognize and represent different variables affecting investment risk and eventually predict the cost of the project. The general practice in cost analysis is to divide the project into smaller cost variables and probabilistically estimate the uncertainty of each item which ultimately assists in the final calculation of project cost. However, dependencies among these items should also be considered otherwise the accuracy of the cost estimation is jeopardized.

A project success is dependent on the careful calculation and organization

The model in this paper suggests the handling of the correlation issue by modeling the uncertainty and indicators which affect cost items. Introduction

Project managers use cost analysis in order to make estimates about the cost of the projects. The difficulty in this type of analysis is that, when a project becomes complex, there are a lot of items that affects the cost. Cost analysis also includes cost uncertainty analysis which helps decision makers to understand not only the potential investment risk but also the nature of risks for a particular project and help to understand possible measure to mitigate such risks.

The trouble in project cost uncertainty is that the interdependency between the cost items or the effect of cost changes of one cost item to another is rarely considered. Without considering such factors, there is always a risk of the over exceeding the initial cost estimation of a project. Such risk also creates other risks like times delays and performance issues (Elkjaer, 2000) (Wang, Wang, & Lai, 2008).The success of a project has generally been associated with the achievement of the goal estimation of time, cost and quality of the project (Atkinson,

It hasfrequently been used in various fields like medical diagnosis, forecasting, automation, and productioncontrol (Heckerman, Mamdani, & Wellman, 1995). A Bayesian belief network is constituted of both quantitative as well as qualitative parts (Van Der Gaag, 1996).All projects are unique to each other by nature (Project management institute, 2013).Due to this reason; the data and facts obtained from one project may not always be relevant or if relevant may not always be applied to others due to their variations. Often the data collection from long-term projects is timely and costly. However, experts and their opinionsare often obtainablewhen required which can be beneficially used as a source of information (Shepherd, Hubbard, Fenton, Claxton, Luedeling, & de Leeuw, 2015). This is where BN comes as a handy tool to take advantage of this expert knowledge with combination of available facts and information to derive necessary prediction whenever