A long term expected risk estimation of maritime accidents through Markov chain approach and probabilistic risk matrix

• Published in: Asian Journal of Shipping and Logistics Vol. 39; no. 3; pp. 1 - 12
• Main Authors: Kim, Dong Jin, Sur, Ji Min, Cho, Hyeon U.
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.09.2023; Elsevier; 한국해운물류학회
• Subjects: A Markov chain; Consequence; Frequency; Probabilistic risk matrix; Risk prediction; Uncertainty; 무역학; Frequency; Consequence; Risk prediction; A Markov chain; Uncertainty; Probabilistic risk matrix
• ISSN: 2092-5212
• DOI: 10.1016/j.ajsl.2023.04.002

Among many risk assessment techniques, qualitative, semi-quantitative or quantitative, risk matrix is a common tool to assess risk by allocating frequency and consequence of an accident to one of the pre divided frequency and consequence categories. However, since there is no standardized way to define these categories, risk matrix with its strength of being straightforward to use and easy to interpret has inherent uncertainties including objective categorization determination and consistency of decisions on frequency and consequence among different users, which usually results in multiple risk outcomes leading to incorrect conclusion about decision making on risk initiated from accidents. The purpose of the study is to predict a long term risk of maritime accident using a 5 × 5 probabilistic risk matrix where each category of frequency and consequence is probabilistically estimated by a Markov chain model. The proposed method to calculate accident risk is illustrated using maritime accident data over 2016–2020 years. The findings are that the most probable frequency and consequence ranges of maritime accidents will be between 182 and 235 with probability of 0.3878 and between 6.8 and 11.6 fatalities with probability of 0.3791, respectively. The expected risk value was computed as 4.6506 on a scale of 2–10. For the validation of the proposed method 90 %, 95 %, and 99 % confidence intervals were constructed which were shown to contain the predicted risk value. The probabilistic risk matrix with Markov chain approach can be applied to predicting risks in different fields of industries.