将随机变系数的混料模型转化为混合效应模型,从而可以得到固定效应的信息矩阵及其逆矩阵,同时证明如果一个设计在固定效应部分是V-最优,则这个设计在整个模型也是V-最优.给出在V-最优准则下的一阶q分量和二阶2分量随机变系数混料模型的V-最优设计.
This paper studies mixture model with random coefficients that can be transformed into a mixed effects model. We obtain the information matrix and its inverse matrix of fixed effects. We prove that mixture design is V-optimal in the whole model if it is V-optimal in the fixed effect part. The V-optimal designs for firstorder q components and second-order two components' mixture model with random variable coefficients are given.