2017/09/07 by Radoslav Harman, Harman, Radoslav, Maryna Prus +1
Decision Sciences · Engineering · #62K05 #Computation (stat.CO) #FOS: Computer and information sciences #Manufacturing Process and Optimization #Optimal Experimental Design Methods #Technology Assessment and Management
paper · pdf · doi:10.48550/arxiv.1709.02317
openalex publication_date 2017/09/07 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
We consider the problem of computing optimal experimental design on a finite\ndesign space with respect to a compound Bayes risk criterion, which includes\nthe linear criterion for prediction in a random coefficient regression model.\nWe show that the problem can be restated as constrained A-optimality in an\nartificial model. This permits using recently developed computational tools,\nfor instance the algorithms based on the second-order cone programming for\noptimal approximate design, and mixed-integer second-order cone programming for\noptimal exact designs. We demonstrate the use of the proposed method for the\nproblem of computing optimal designs of a random coefficient regression model\nwith respect to an integrated mean squared error criterion.\n