Required characteristics of statistical distribution models for life cycle cost estimation

Life cycle cost (LCC) is used as the decision objective in many engineering problems such as project selection and bidding. An important nature of LCC estimation is uncertainty. Therefore, cost elements and system cost are usually represented by statistical distributions. A number of statistical models have been used for modeling LCC in the literature. It appears to be lack of sufficient attention to their appropriateness. This paper classifies the models into two categories: one for modeling cost elements and the other for modeling sub-system and system costs. A set of required characteristics for each type of models is identfied. The paper finds that most of the existing models do not fully satisfy the identified requirements. Therefore, two new models are proposed, and their appropriateness is illustrated by a numerical example.