Abstract:

Quality has been used as a decisive factor in choosing products and services. As a result， control chart has been widely researched and applied as an effective quality control tool. When the measurement of a specific quality characteristic is expensive and time consuming， common control chart methods are often difficult to play a effective role in process control. For this reason， n{p}_x chart is proposed as an attribute control chart applied to monitor mean shifts of a variable， which has the advantages of simple measurement and cost saving， but its ability to monitoring abnormal shifts is often not strong. In order to improve the monitoring efficiency， a dynamic n{p}_x chart design method is proposed in this paper based on dynamic control chart method. As an accurate assessment of monitoring efficiency， average product length of dynamic n{p}_x chart is calculated by probabilistic method. Based on Taguchi quadratic quality loss function， average quality cost per unit of product is proposed and calculated to evaluate quality cost of dynamic n{p}_x chart. Taking average product length in control as constraint， average product length out of control and average quality cost per unit of product as objective functions， a multi-objective optimization design model of this chart is established. A numerical example of the model is calculated by non-dominated sorting genetic algorithm （NSGA） and is used to compare with static \mathrm{n}{\mathrm{p}}_{\mathrm{x}} chart， variable parameters （VP） chart， exponentially weighted moving average （EWMA） chart， the results show that the optimal design method proposed in this paper is superior to other control chart methods. The research will further enrich the theory of control chart optimization design， and provide an easy-to-use and efficient quality control method for quality management practice.

Key words: dynamic control chart, n{p}_xchart')">n{p}_xchart, average product length, multi-objective optimization