摘要
本文在Fedorov算法的基础上,引入了Schmidt初始化策略,对原算法进行了改进。更进一步地,结合最小体积闭包椭球问题的理论、支持向量机中序列最小最优化(Sequential minimal optimization,简记为SMO)算法思想[2],和Schmidt初始化策略给出了D-最优设计一个新的数值算法——SMO更新算法,并对其复杂度进行了分析。
In this paper,based on the Fedorov method,the original algorithm was improved by introducing the initial strategy of Schmidt.Furthermore,combining the minimum volume eppipsoid closure theory,the support vector machine（SVM） minimum sequence optimization（Sequential minimal optimization,referred to as SMO） algorithm thought,and the initial strategy of Schmidt,we advance a new method for D-optimal experimental design——SMO updating algorithm,and its complexity is analyzed.
出处
《长春理工大学学报（自然科学版）》
2012年第3期86-88,92,共4页
Journal of Changchun University of Science and Technology(Natural Science Edition)
