Quality Based On Complexity Theory, Systems Analysis

Triggered by the improvement of the science and technology and the market demand, quality system increasingly becomes giant and complex and has evolved into a typical complex system. Thus some problems, which can't be solved with previous theories and methods, emerged. As a new stage of system science, complexity science studies systems and their complexity. It will have great influence on the reaserch of quality system in the 21 st century.In this dissertation, we focus on the quality system. The theories and analytic methods of complexity science are applied to establish quality improvement model, customer satisfaction model and analyze the evolution of the product market network. These studies involve the fundamental theories, methods and application in many fields, such as quality management, complexity theory and so on.The main contents in this paper can be summarized as follows:Firstly, based on the review of the development of the quality ideas, quality management and quality system, we analyze the characteristics of the quality system expounding that the quality system is a typical complex system and the product quality improvement is a typical complex adaptive system.Secondly, NK model is used to depict the quality system and study its innovation. Three quantitative description forms of NK model are presented in this paper based on the idea of NK model.Thirdly, we improve the GMDH (Group Method of Data Handling) and apply it to rubber quality improvement.Fourthly, the customer satisfaction model is a nonlinear model with dichotomous dependent variables. This paper presents a logistic simultaneous equations model with latent variables. Factor analysis is used to extract the latent variables which affect the satisfaction and loyalty. Logistic regression can be done with the latent variables and the two dependent variables--the customer satisfaction and loyalty, then a recursive simultaneous equations model are established. Finally, an example about community satisfaction is given to validate the model above.Finally, based on the introduction of the BA growing networks, an extended BA model is proposed in this paper. The equivalence of master equation and Polya process model has been proved by the theory of probability. According to this, we discuss some kinds of probability functions, which depict the customers' choice, including linear probability model, Logistic model, and OC curve model. With these models and the analysis of their fixed points, the influence on the macroscopical behavior of the market is investigated. At last, several examples are given to illustrate and validate the conclusion of this paper.