Marine and Offshore Safety Assessment by Incorporative Risk Modeling in a Fuzzy‐Bayesian Network of an Induced Mass Assignment Paradigm

Abstract

The incorporation of the human element into a probabilistic risk-based model is one that requires a possibilistic integration of appropriate techniques and/or that of vital inputs of linguistic nature. While fuzzy logic is an excellent tool for such integration, it tends not to cross its boundaries of possibility theory, except via an evidential reasoning supposition. Therefore, a fuzzy-Bayesian network (FBN) is proposed to enable a bridge to be made into a probabilistic setting of the domain. This bridge is formalized by way of the mass assignment theory. A framework is also proposed for its use in maritime safety assessment. Its implementation has been demonstrated in a maritime human performance case study that utilizes performance-shaping factors as the input variables of this groundbreaking FBN risk model.

Human Error Contribution in Collision and Grounding of Oil Tankers

“…These two questions indicate the necessity for the improvement of the usual HRA methods. ( 13 ) Some articles purposed to model the system through the Bayesian network, ( 14,15 ) to improve the individual representation and interrelations. Thus, it is expected that the application of this tool will compensate for part of the flaws left by the use of binary trees.…”

Bayesian networks for maritime traffic accident prevention: Benefits and challenges

“…One alternative for this would be the integration of BNs with other approaches to uncertainty representation than probabilities. For example, Li et al (2012) established a fuzzy BN network approach for the quantification of organizational influences in human reliability analysis context, and Eleye- Datubo et al (2008) proposed a similar model for the maritime safety domain. Various ways to elicitic fuzzy knowledge from experts have been presented, e.g.…”

Probabilistic modeling of navigation bridge officer's behavior

“…Indeed, published work provides evidence that the incorporation of human factors in marine risk assessments has gained increased popularity. The adopted approaches can vary from integrating human reliability methods with artificial intelligence techniques [5], analyzing the role of human error with Bayesian network [6,7,8,9] and extending it with fuzzy techniques [10], or utilizing classical probability theory [11].…”