Maximum Likelihood Parameter Determination Method for Complex System Modeling
Models are often used to characterize a complex system in analysis problems. In modeling process, it is very difficult that model parameters are determined. So, a maximum likelihood parameter determination method for complex system modeling is presented to solve this problem. Therefore, a merge method is adapted to achieve the model parameter for complex systems. The presented method is a kind of merge way and a statistics mode. The parameter values are obtained by having datum is summarized. In this text, the distribution function of unit life is established according to their probability properties. The expressions of the unit failure probability are gotten respectively. Because electromechanical system lifecycle always follows the Weibull distribution, and there are these limitations of small sample and incomplete data, the exponential distribution function is applied as a special way to determine parameter values. Then an extrapolation mode is adapted for the parameter computing. Finally, an example is explored to illustrate the proposed methods. This result is shown that presented methods are effective and feasible. And this method can be widely applied in model parameter determination for another complex system.
maximum likelihood complex system Weibull distribution small sample incomplete data
Rui JIN Zhong HAN
Department of Computer Science and Engineering Henan Institute of Engineering Zhengzhou, China Institute of Manufacturing Systems and Quantity Engineering Xi’an Jiaotong University Xi’an, China
国际会议
西安
英文
369-372
2011-06-17（万方平台首次上网日期，不代表论文的发表时间）