摘要
在分析模糊Petri网推理机制的基础上,将优化算法ACA(Ant Colony Algorithm)引入至FPN(Fuzzy Petri Net)的学习能力问题中。针对一知识库系统的具体实例,探讨该算法在FPN学习能力问题中的具体实现,并结合传统优化算法对比分析了它们各自的特点和性能优劣。仿真实验表明,ACA算法整体性能最佳,训练出的参数正确率较高,且所得的模糊Petri网具有很强的泛化能力和自适应功能。
Based on analysis on the Fuzzy Petri Net（FPN） reasoning mechanism,the paper introduces Ant Colony Algorithm（ACA） into the FPN learning capability problem,discusses the actual implementation of the algorithm in the FPN learning capability problem with a detailed instance of a knowledge base system,then by combining conventional optimisation algorithms compares and analyses their characteristics, functional pros and cons respectfully.Emulation experiment illustrates that ACA wins for its overall performance not only for its accuracy rate of its trained parameters,but also for its enormously powerful generalisation and self - adaptability capabilities.
出处
《计算机应用与软件》
CSCD
2010年第11期127-130,共4页
Computer Applications and Software
基金
湖南省教育厅自然科学基金资助项目(01JJY2061)
湖南省教育厅科研基金资助项目(01C306)
关键词
模糊PETRI网
遗传算法
BP算法
蚁群算法
克隆选择算法
Fuzzy Petri Net（FPN）
Genetic algorithm
BP algorithm
Ant colony algorithm
Clone selection algorithm
