A study on comparison of screening designs on optimality criteria

Statistical experimental design method is a quality improvement and improvement method which has wide usage area in many active sectors such as industry, engineering, manufacturing especially in recent years. Experimental design methods make a significant contribution, especially in the design of a new final product, in the planning, improvement and development of the production phases of this product. In today's competitive market conditions, the primary aim of the companies is to produce high quality products with low expenses. Statistical experimental design methods are widely used in almost every field of industry today. In the first part of this thesis, the definitions of the experimental design, its historical process, the basic principles on which it is based, its application areas and sectors and its application process are discussed. In the second chapter, the factorial designs were introduced, firstly the design types used in the first order models and then the second type models. In the third chapter, Taguchi type designs and the orthogonal arrays of Taguchi are discussed. In the fourth chapter, the design types used as statistical elimination method are mentioned. In this section, 2-level factorial designs, Plackett-Burman designs and descriptive screening designs are explained. In the last chapter, the criteria of optimality used to select the optimal design are emphasized. In the scope of this study, one of the optimality criteria, A optimality, D Optimality and E Optimality were discussed. Again in this section, an application for comparison of screening designs over the criteria of optimality is given.