本計畫旨在研究當有先前資訊可供使用時,如何找出A最適貝氏設計和.GAMMA.-minimax設計之充分條件。在本報告中同時並將針對兩個先前母數來探討A最適貝氏設計和.GAMMA.-minimax設計之穩健性,以及探討平衡處理集區設計之穩健性。表列之穩健最適設計和高效率之設計,亦將在本計畫之最後部分給出。 The problem of comparing a set of .nu. test treatments simultaneously with a control treatment when k>.nu. is considered. Following the work of Majumdar (1992), we use exact design theory to derive Bayes A-optimal block designs and optimal .GAMMA.-minimax designs for the one-way elimination of heterogeneity model. We also observe that Bayes optimal designs are very robust against the departures from the given prior distribution. Tables of robust optimal designs and highly efficient designs with a general error prior assumption are provided at the end of this report.
