Measurement of Fuzzy Membership Functions in Construction Risk Assessment

Abstract

Fuzzy set theory (FST), especially its concepts of linguistic variables and fuzzy membership functions (FMFs), has been frequently used in construction risk assessment due to a lack of historical data and reliance on expert opinions. However, FMFs have to be properly defined to leverage the ability of FST in dealing with vagueness, imprecision, and uncertainty in expert opinions. This study aims to identify risk variables and their associated FMFs used in relevant literature and develop a process of measuring FMFs. The research methodology involves conducting a systematic review, developing a process for defining FMFs, and validating the proposed process through an illustrative case study with hurricane storms as a risk event. Results showed that only 11 risk factors represented 77% variables used in 131 articles that were found relevant for further analysis. Most studies (88%) arbitrarily assigned FMFs to their risk variables despite the fact that defining FMFs is one of the critical steps for the successful use of FST-based methods. Membership functions have to be measured with relevant data, including those provided by experts. To fill this knowledge gap, a step-by-step process is developed based on the modified horizontal approach and applied in the case study to appropriately determine the FMFs of the risk variables. Separate empirical hurricane-related public data were used to verify the determined FMFs. This study contributes to the body of knowledge by providing standardizing FST-based methods for risk assessment and measuring the FMFs of risk variables. The study also facilitates the application of FST-based methods in actual construction projects and operations.