摘要
针对运用传统评估算法评估复杂且耦合性强的操作及一些特殊操作时所得的评估结果与实际差别很大的情况,将专家系统应用于船舶训练模拟器操作考核评估中,以有效解决传统考核评估中存在的诸多问题。通过分析船舶训练系统操作管理知识的特点,提出知识的层次组织形式。采用基于产生式规则和模糊产生式规则的知识表示方法,将系统知识分为简单型知识表示、顺序过程型知识表示、关联型知识表示和控制型知识表示,并根据实际情况灵活确定知识表示的类型,以达到最好的知识表示效果。由于船舶训练系统操作考核评估问题搜索空间较小,因此系统搜索策略采用深度优先搜索法,以保证其快速性。
When traditional examination algorithms are used for evaluating complex and strong coupled operations or some special operations,its output marks often deviate from the actual performances noticeably. For solving the problem,introduces the expert system to the ship training simulators. The characteristics of operation management knowledge of ship training systems are analyzed and a hierarchical structure is constructed to organize the knowledge. The system knowledge is divided into simple knowledge,sequential process type knowledge,association type knowledge and control type knowledge according to the knowledge production rules and fuzzy production rules with certain flexibility in representation of some kind of knowledge for better effect. The limited search space of the assessment problem allows the system uses a depth-first search strategy which ensures rapid response of the system.
出处
《中国航海》
CSCD
北大核心
2015年第4期59-63,共5页
Navigation of China
基金
国家自然科学基金(51179196)
关键词
专家系统
训练模拟器
知识表示
模糊产生式规则
推理机
expert system
training simulator
knowledge representation
fuzzy production rule
inference engine
