A-OPTIMAL AND EFFICIENT COMPLETELY RANDOMIZED DESIGNS IN DIALLEL CROSS EXPERIMENTS FOR COMPARING TEST TREATMENTS WITH A CONTROL

• Main Authors: Jo-Yun Lin, 林若鋆
• Other Authors: Yung-Feng Hsu
• Format: Others
• Language: zh-TW
• Published: 2006
• Online Access: http://ndltd.ncl.edu.tw/handle/33f9hg

碩士 === 銘傳大學 === 應用統計資訊學系碩士班 === 94 === Experimental designs for diallel crosses are commonly used in plant and animal breeding experiments to study the genetic properties of inbred lines. Many articles about the heredity and mates have been studied for several decades. The study of optimal designs for diallel cross experiments has been receiving considerable attention recently. In this thesis, we focus on finding A-optimal and efficient Type-S designs in diallel cross experiments for comparing test lines with a control line under the model of completely randomized designs. We use the definitions of A-optimal and A-efficient designs and the computer program to find A-optimal and efficient Type-S designs. The families of A-optimal designs are derived when the number of test lines is two, three and four. Also the families of efficient Type-S designs are found out when the number of test lines is between four and ten.