ABSTRACT
This paper will conduct a comprehensive study, the introduction of the concept of quality loss function and preventive maintenance, aiming to establish and optimize a more consistent with the economic model of the actual process management. According to the sample data of normal distribution, based on two principles of quality loss function and the preventive maintenance, considering the control costs of quality control stages, making use of the Duncan classic model to design economic model of CUSUM chart, so that minimizing the risk cost of a process cycle. Then, the average running length and the average alarm time of the CUSUM control charts are calculated by using the Markov chain method (Elliott et al. 1995), and the algorithm implementation process is given. This paper use genetic algorithm to optimize decision parameters of the model of sample size (n), sampling interval (h), reference value (K), warning width (w), decision distance (H) and deduce sensitivity analysis of decision parameters, improving the design accuracy of the parameters in the risk decision of quality management.
