次序試驗的設計與分析

Order-of-addition (OofA) experiments are designs where the output of a response depends on the addition order of adding $m$ different components and a series of experiments was conducted to evaluate different orders of addition. It is often unaffordable to test the all possible combinations of components which requires $m!$ runs (for example, 10! is about 3.6 millions), and the design problem arises to choose a subset of orders for comparison. In this project, we will review the statistical model for analyzing data from such experiments and study associated design problems. In this project, we will focus on the nearest neighbor effects for the analysis the order-of-addition experiment and the construction of effective order-or-addition experiments is studied.