DoE is best characterized as making several assumptions about the design or the process being studied, quantifying these assumptions by the choice of factors and levels, and then running experiments to determine if these assumptions are valid. It a mix of several tools that has been developed to optimize performance, based on statistical analysis, significance tests, and error calculations.
Improving the process capability requires the concurrent efforts of both product and process designers. Product designers should increase the allowable tolerance to the maximum that will still permit the successful functioning of the product. Process designers should center the process to meet the specification target and minimize the variability of the process. DoE is a tool that can help with both of these goals.
DOE uses statistical experimental methods to develop the best factor and level settings to optimize a process or a design. Some of the statistical methods have been simplified by the use of specialized software analysis packages. In many cases， the engineers responsible for the process or design can perform the necessary steps to conduct the experiments from knowledge presented in this or other books on DoE, or after taking minimum training in the techniques of DoE, perhaps with the assistance of a statistician. In addition，the technical knowledge of the basic science or technology necessary for optimizing a design is not critical. Neophytes can optimize product and process design just as well as experienced engineers using DoE techniques.