摘要
在遗传算法基础上提出了多目标测试点确定问题的新的解决方案。一方面由设计者根据对关键元件的关注程度,提高其相关测试点选中权值或直接预置隔离,另一方面在此基础上算法搜索出目标综合要求最高的测试点集来。为了把处于高适应度前沿的优秀编码之间的距离拉大,采用了指数方法。另外,为提高搜索的准确性,提出歧视规则的变异思路。最后通过对应用实例的仿真结果验证了算法的合理性和有效性。
A new method which solved the test point＇s problem of multi-objective was introduced based on the interaction genetic algorithms. In the course of calculating, the deviser may raise the key-component＇s hitting weight or straight insulate them in advance, according to the attention degree of them, At the same time, the algorithms seeked out the test point set in which synthetical requirement is the upmost about objective based on the former, The method introduced the index measure for increasing the distance of excellent codings which have higher fitness than the others. In addition, a mutation way known as discrimination rule for raising the veracity of the search was put forward. Finally, the simulation results of an applied example demonstrate that the algorithms is rational and effective.
出处
《系统仿真学报》
EI
CAS
CSCD
北大核心
2006年第6期1469-1472,共4页
Journal of System Simulation
基金
国家自然科学基金(50275125)
航空科学基金(04I53068).
关键词
测试点
遗传算法
歧视规则
多目标
交互式
数据融合
test point
genetic algorithms
discrimination rule
multi-objective
interaction
data fusion
