Risk assessment model with probabilistic linguistic fuzzy inference methods for maritime piracy crime and applications

The surge in maritime piracy is becoming a deterrent to international seaborne trade. More attention should be attached to piracy risk analysis and prevention. The complex environments engendering piracy crime make it more difficult to conduct a quantitative judgment. Most of the existing research focuses on the risk assessment of piracy based on historical piracy activities. The dynamic evolution of piracy is not explored in these evaluation models. To solve the piracy evolution pattern analysis and risk control in the uncertain environment, we first propose a novel probabilistic linguistic Markov model to analyze and predict the risk states of piracy accidents. The basic definition of random states and the Markov process in the probabilistic linguistic environment are discussed. Next, we propose the probabilistic linguistic Bayesian network which not only depicts the causal relationships of the risk factors but also provides an efficient tool to calculate the transition probabilities in the Markov model. Then the comprehensive inference risk analysis methods are developed to conduct the risk assessment in the maritime piracy crime. The assessment results can effectively support the decision-making of risk control. A practical case study is conducted to validate the above methods. Sensitivity analyses and method comparisons also illustrate the performance of the proposed risk analysis model.