摘要
针对齿轮箱流体动压轴承多目标优化问题,提出了一种基于混合离散变量多目标优化设计求解方法,该方法应用灰色综合关联度将多目标优化问题转化为单目标优化问题,再用改进微粒群算法求解,开发了混合离散变量优化的灰色改进微粒群法程序。该方法能合理地处理优化设计中混合离散变量的取值问题,通过对标准微粒群算法改进,引入动态罚函数,构造一种新的适应函数,算法具有很强的全局寻优能力。优化设计实例表明,该算法对优化设计问题的特性无特殊要求,具有较好的普适性,而且程序运行可靠,全局收敛能力强。
Aimed at the multi-objective design problem of hydrodynamic sliding bearing in gearbox,a general optimization method was introduced to high dimension multi-objective optimization with hybrid discrete variables.Based on the proposed method the multi-objective was transformed into single object with synthetic degree of grey incidences,and the optimal solutions were found by improved particle swarm optimizer(PSO).The grey improved PSO algorithmic program for the multi-objective optimization with hybrid discret...
出处
《润滑与密封》
CAS
CSCD
北大核心
2008年第4期95-98,105,共5页
Lubrication Engineering
基金
湖南省“十一五”重点建设学科资助项目(湘教通[2006]180)
湖南省自然科学基金项目(07JJ3093)
湖南省科技厅计划项目(2007FJ3030、2007GK3058)
关键词
动压滑动轴承
灰色综合关联度
微粒群算法
适应函数
动态罚函数
hydrodynamic sliding bearing
synthetic degree of grey incidence
particle swarm optimizer
fitness function
dynamic penalty function
