基于ANFIS机理的船舶溢油损害评估模型

Vessel Oil Spill Damage Evaluation Model with Adaptive Neural-Network-Based Fuzzy Interference System

考虑到船舶溢油事故随机性大、影响因素复杂,通过层次分析法确定溢油量、海域情况、清污情况和环境敏感度作为船舶溢油损害影响因子集。基于对自适应神经模糊推理系统(Adaptive Neural-network-based Fuzzy Interference System,ANFIS)建模机理的研究建立非线性模糊模型,并结合国际油污赔偿基金公布的事故数据进行溢油损害的仿真与验证。分析结果在合理的误差范围内,证实了该模型在船舶溢油损害评估中的可行性。

By an analytic hierarchy process,the oil spill quantity,clean-up response,oil spill area and the environment sensitivity are identified as the influencing factors of the damage of oil spill accidents,which possess strong randomness and complexity.The Adaptive Neural-network-based Fuzzy Interference System（ANFIS） is introduced,and a nonlinear fuzzy model for evaluating oil spill damage is established.For verifying the model,the simulation is performed and based on the oil spill data assembled by the International Oil Pollution Compensation Fund（IOPCF） and the simulation results are checked,which indicates that the estimates are reasonably accurate,implying the feasibility of the ANFIS evaluation model in dealing with vessel oil spill compensation.