Optimal Efficiency Balanced Designs And Their Constructions Under The Randomization Model

Randomization and local control are the basic principles which experimental de-sign is based on. Randomization can decrease the experimental error that influence the analysis of experimental result. Local control can reduce the negative affects because of nuisance factors and make experiment more effective. In this paper, we discuss ef-ficiency balanced design and its statistical properties under the randomization model. We give the optimal efficiency balanced designs and their constructions. In the second chapter, we give the basic terminology and preliminaries. In the third chapter, we dis-cuss statistical properties of the treatment effect under randomization model and its three submodels in detail. We obtain the optimal efficiency balanced design in a class of non-proper binary efficiency balanced designs which have the some parameters in the next chapter. We also get the statistical relationship between canonical efficiency factors and basic contrast. Then, we construct two kinds of optimal efficiency balanced design. Finally, we sum up the full text of some of the issues raised for further study.