仇光伟 胡以怀 郭晓亮

摘要：为更好地提高现有轮机模拟器实操评分的准确性，参考国内轮机模拟器自动评分系统的原理，提出一种多级模糊综合评价方法。该方法依据中华人民共和国海事局在《海船船员适任考试与评估大纲》中规定的实操技能要求，建立评价指标体系。使用层次分析法（analytic hierarchy process，AHP）确定各指标权重，选择适宜的隶属度函数模型确定各指标的隶属度。实时采集船员在实操训练过程中产生的数据，依据专家系统进行模糊评价。经专家验证取得了理想的结果。模型提高了实操评分的准确性、公正性及其评价效率，可为海事部门的考核提供有力的支撑。

关键词： 轮机模拟器; 实操训练; 自动评分; 多级模糊综合评价; 层次分析法（AHP）

中图分类号： U676.2    文献标志码： A

Automatic scoring method and realization of SMSC marine engine

room simulator practical training

QIU Guangwei， HU Yihuai， GUO Xiaoliang

（Merchant Marine College， Shanghai Maritime University， Shanghai 201306， China）

Abstract： In order to improve the accuracy of practical training scoring of existing marine engine room simulators， referring to the principle of the domestic marine engine room simulator automatic scoring system， a multi-level fuzzy comprehensive evaluation method is proposed. In the method， an evaluation index system is established based on the practical operational skill requirements stipulated by the Maritime Safety Administration of the Peoples Republic of China in the Shipboard Crews Qualification Examination and Assessment Outline. The analytic hierarchy process （ AHP） is used to determine the weight of each index， and the appropriate membership function model is selected to determine the membership degree of each index. The data produced by the crew in the practical training process are collected in real time， and the fuzzy evaluation is performed according to the expert system. The ideal result is obtained by expert verification. The model improves the accuracy， fairness and evaluation efficiency of the practical training scoring， and can provide strong support for the assessment of maritime departments.

Key words： marine engine room simulator; practical training; automatic scoring; multi-level fuzzy comprehensive evaluation; analytic hierarchy process （AHP）

0 引 言

據交通运输部新闻办公室发布的《2017年中国船员发展报告》显示，截至2017年底我国共有注册船员148.3万余人，其中海船船员70.9万余人。在船员队伍总体保持稳定增长的同时，现代船舶也在向着大型化、专业化的方向发展，这对船员的管理水平和实操水平提出了更高的要求[1-2]。

为避免船员实操人工评估的主观性、延时性，需要设计轮机模拟器的自动评分系统。目前自动评估算法包括：基于专家系统评分规则的模糊评估算法[2]、采用专家评价法和统计评判法定权的模糊综合评价法[3]、分解综合法结合模糊综合法[4]、采用专家综合测评方式定权的模糊综合评价法[5]等。每种方法都有自己的优劣，虽然框架各异，但其核心问题均为怎样合理地确定各因素的权重和隶属度。……