摘要
为解决遗传算法存在的早熟、收敛全局最优解难题,提出适应度标定问题;定义相似度概念,剔除相似性个体。提出了CPU模板概念,解决了外围模块与CPU的耦合关系,提出了仪表组成模块的性能参数与仪表的性能参数之间的映射关系。改进遗传算法解决了传统的遗传算法存在问题,并且解决了穷举法进行模块组合的效率问题,基于此算法的推理机制能够做到设计新仪表,并且新仪表能基本符合用户要求。
In order to avoid prematurity and convergence out of optimized point for simple genetic algorithms,a fitness normalization formula was introduced.The similarity was defined to eliminate the similar individual.CPU template conception was introduced for solving the problem about Coupling relations between CPU and module,also put forward mapping function about performance parameter between component and instrument.(Improved) genetic algorithms can solve the time efficiency problem which exhaust algorithms have.The reasoning engine base on(genetic) algorithms can design new instrument which can meet the requirement of user basically.
出处
《机电工程》
CAS
2006年第8期5-7,14,共4页
Journal of Mechanical & Electrical Engineering
关键词
专家系统
遗传算法
模板
expert system
genetic algorithms
template
