A batch-by-batch input experiment design method based on d-optimal design

Abstract

The invention provides a batch-by-batch investment test design method based on D-optimum design. The method comprises the following steps of 1, determining basic test information; 2, giving the D-optimum design scheme Epsilon <n><min> of the minimum number of test times n<min> by using a D-optimum design method; 3, determining batch-by-batch investment design batch nodes until the maximum number of test times n<max> allowed by the test resources is reached; 4, performing batch-by-batch test scheme design; 5, determining the optimum number of test times on the basis of G-efficiency; 6, performing the test and model verification; performing the batch-by-batch investment test according to the test scheme; and performing the model inspection; and 7, performing regression fitting to obtain unknown parameters in a relationship model y, and determining the optimum factor level combination. The method provided by the invention solves the test design problems under the conditions of test region irregularity, high factor dimension, test time limitation and the like, and has the advantages that on the basis of ensuring the test design efficiency, the influence caused by some uncertainty in the design engineering is relieved; the realization is easy; the use by engineers and technicians is convenient; and good application values are realized.

Description

technical field [0001] The invention relates to a batch-by-batch input test design method based on D-optimal design, which can effectively solve the constraints between the number of tests in traditional test design and design optimality, and ensure that batch-by-batch tests have a good design optimality. Excellent properties, suitable for product design, manufacturing, quality control and process optimization and other related technical fields. Background technique [0002] A reasonable experimental design scheme can fully reveal the impact of each influencing factor on the target response under limited time, cost and other resource conditions, obtain the optimal target response result under a limited number of experiments, and seek the optimal solution for improving product quality. combination of process parameters, etc. [0003] Broadly speaking, all experimental designs are resource-constrained designs. The purpose of experimental design is to obtain as much system or ...

AI Technical Summary

Problems solved by technology

[0007] The present invention aims at the traditional D-optimal test design, the number of tests cannot be flexibly adjusted according to actual needs, once the test starts, the number of tests cannot be changed, otherwise it will affect the optimality of the design, and provides a D-optimal design based The batch-by-batch test design method can ensure that the test plan has better optimal properties when the number of tests changes, establish a mathematical model close to the actual product processing or production process, and conduct accurate data analysis and prediction, so as to determine The best combination of factor levels, optimized design and process, has important theoretical significance and urgent practical needs

Images