A review of artificial intelligence based risk assessment methods for capturing complexity-risk interdependencies

Abstract

Purpose In the past decades, artificial intelligence (AI)-based hybrid methods have been increasingly applied in construction risk management practices. The purpose of this paper is to review and compile the current AI methods used for cost-risk assessment in the construction management domain in order to capture complexity and risk interdependencies under high uncertainty. Design/methodology/approach This paper makes a content analysis, based on a comprehensive literature review of articles published in high-quality journals from the years 2008 to 2018. Fuzzy hybrid methods, such as fuzzy-analytical network processing, fuzzy-artificial neural network and fuzzy-simulation, have been widely used and dominated in the literature due to their ability to measure the complexity and uncertainty of the system. Findings The findings of this review article suggest that due to the limitation of subjective risk data and complex computation, the applications of these AI methods are limited in order to address cost overrun issues under high uncertainty. It is suggested that a hybrid approach of fuzzy logic and extended form of Bayesian belief network (BBN) can be applied in cost-risk assessment to better capture complexity-risk interdependencies under uncertainty. Research limitations/implications This study only focuses on the subjective risk assessment methods applied in construction management to overcome cost overrun problem. Therefore, future research can be extended to interpret the input data required to deal with uncertainties, rather than relying solely on subjective judgments in risk assessment analysis. Practical implications These results may assist in the management of cost overrun while addressing complexity and uncertainty to avoid chaos in a project. In addition, project managers, experts and practitioners should address the interrelationship between key complexity and risk factors in order to plan risk impact on project cost. The proposed hybrid method of fuzzy logic and BBN can better support the management implications in recent construction risk management practice. Originality/value This study addresses the applications of AI-based methods in complex construction projects. A proposed hybrid approach could better address the complexity-risk interdependencies which increase cost uncertainty in project.

Analysis of Cost Overrun and Schedule Delays of Infrastructure Projects in Low Income Economies: Case Studies in Ethiopia

“…Data analysis is the process of analyzing, testing, and connecting a number of qualitative and/or quantitative data to address specific objectives and research questions [48]. For this study, the data collected from questionnaire survey is analyzed using popular statistical analysis tool, the Statistical Package for Social Science (SPSS), version 23.…”

Exploring Technical Decision-Making Risks in Construction Megaprojects Using Grounded Theory and System Dynamics

“…Meanwhile, ANP and Bayesian network (BBN) can be used to explore the interactions among risks and evaluate the risk state quantitatively, while they require large amounts of data. Wu et al quantify the risk level of a subway station construction using fuzzy ANP via the synthesis of weight matrices, which requires much more computation for pairwise comparison between risk factors [46,47]. BBN performs excellently to model complex relationships among risks on the bases of the conditional probabilities of the nodes [48].…”

Prioritization of interdependent uncertainties in projects

“…Although each modeling technique has its own merits and limitations, BBNs have gained much popularity due to their robust theoretical framework and the ability to capture uncertainty and update beliefs upon the availability of new information (Afzal et al, 2019). These unique features of BBNs make it a suitable framework for mapping project uncertainties as a project uncertainty network.…”