A first-order algorithm for the A-optimal experimental design problem: a mathematical programming approach

被引：0
作者：
Selin Damla Ahipaşaoğlu
机构：
[1] Singapore University of Technology and Design,
来源：
Statistics and Computing | 2015年 / 25卷
关键词：
Optimal experimental design problem; A-optimality;
D O I：
暂无
中图分类号：
学科分类号：
摘要：
We develop and analyse a first-order algorithm for the A-optimal experimental design problem. The problem is first presented as a special case of a parametric family of optimal design problems for which duality results and optimality conditions are given. Then, two first-order (Frank–Wolfe type) algorithms are presented, accompanied by a detailed time-complexity analysis of the algorithms and computational results on various sized problems.
引用
收藏
页码：1113 / 1127
页数：14
相关论文
共 33 条
[1]  
Ahipaşaoğlu SD(2008)Linear convergence of a modified Frank–Wolfe algorithm for computing minimum-volume enclosing ellipsoids Optim. Methods Softw. 23 5-19
[2]  
Sun P(2008)Identification and elimination of interior points for the minimum enclosing ball problem SIAM J. Optim. 19 1392-1396
[3]  
Todd MJ(2013)Modified Frank–Wolfe algorithm for computing minimum-area enclosing ellipsoidal cylinders: theory and algorithms Comput. Geom. 46 494-519
[4]  
Ahipaşaoğlu SD(1973)Sequences converging to D-optimal designs of experiments Ann. Stat. 1 342-352
[5]  
Yıldırım EA(2009)Multiplicative algorithms for computing optimum designs J. Stat. Plan. Inference 139 3947-3961
[6]  
Ahipaşaoğlu SD(2006)Optimal designs for dose-response models with restricted design spaces J. Am. Stat. Assoc. 101 747-759
[7]  
Todd MJ(1974)On the allocation of linear observations Comment. Phys. Math. 44 27-78
[8]  
Atwood CL(1956)An algorithm for quadratic programming Naval Res. Logist. Q. 3 95-110
[9]  
Torsney B(1993)Optimal designs for nonlinear regression models Commun. Stat., Part A - Theory Methods 22 1613-1627
[10]  
Martin-Martin RR(2007)Improvements on removing non-optimal support points in D-optimum design algorithms Stat. Probab. Lett. 77 90-94
1234