Cloudy with a Chance of Fuzzy: Building a Multicriteria Uncertainty Model for Construction Project Delivery Selection

Abstract

The process of choosing a project delivery method is infused with cognitive uncertainties associated with the decision maker. Fuzzy uncertainties arise because of imprecise understanding and subsequent representation of these uncertainties by the decision maker, whereas random uncertainties arise from variance of these imprecisions. Since there are no well-defined rules for spontaneous decisions, in order to be consistently confident in the appropriateness of the chosen delivery method, a structured approach incorporating uncertainty is required. Previously unanswered questions such as (1) what are the sources of uncertainty in project delivery decisions, (2) how do decision makers conceptualize uncertainties, and (3) how the use of existing models increase or decrease uncertainty, were investigated. The answer to these questions revealed that while fuzziness has been accounted for in present procurement theory, the modeling of randomness has been neglected. Therefore, a forward normal cloud model that uses a normal distribution membership function was built to rank decision makers’ preferences. This proposed approach provides a better conceptualization of project delivery decision making, and is shown to be more sensitive in distinguishing alternatives compared on factors, than the interval analytical hierarchy process (AHP) rough-set approach. Survey results showed statistical significance (p<0.05) for the model’s reliability in choosing the preferred delivery method. Overall, construction decision makers can be confident that conflicting project outcomes will less likely result from their recommendations as the proposed cloud model enables clients to perform quantitative calculations incorporating a wider spectrum of uncertainties in their decision.