一种对仿真可信度评估的智能专家系统

摘要

In this paper, the evaluation for simulation system credibility is researched. The traditional evaluation method using human experts suffers from heavy workload, uncertain information, and vague result. Aiming at overco-ming these problems, an intelligent evaluation approach is provided instead of human experts. Based on uncertainty knowledge representation and reasoning, it constructs an intelligent evaluation expert system to carry out evaluation. The paper elaborates the correlation kowledge about uncertainty knowledge representation and measurement, uncer-tainty reasoning, vague information quantitating and data integration. Then this method is demonstrated by the exam-ple of cellular signal transduction network simulation model. The demonstration proves that the simulation model eval-uation using this system is practical, and the result is of impersonality and is believable. It can rival the decision -making process of human experts.%研究对仿真可信度的评估问题.针对传统的人类专家评估方法其工作量大、信息不确定以及结果模糊等问题,提出一种新型的替代专家的评估仿真可信度的方法,以不确定性知识表示和推理为基础,通过构建智能评估专家系统来实现评估.并对不确定性知识的表示与量度、不确定性推理、以及模糊信息的量化和信息的综合等相关知识进行了详细的描述.以细胞信号转导网络仿真模型的评估为实例进行了演示,演示表明:系统对于其仿真模型的评估实用,结果客观,可信.可与人类专家的决策过程媲美.