Improvement Of Sensitivity Analysis’ S Probability Problems

Sensitivity analysis has been widely used in many engineering ?elds. However,it can be also applied to other ?elds, like economics or medical science.There is a particular stimulating threshold for every measurable stimulus response in a product. When the stimulating amount of an individual exceeds the threshold, this individual responds to the stimulation. Otherwise, it gives no response. The stimulating threshold of each individual product is a random variable. The probability distribution of stimulating threshold is called response distribution. In the past, the sensitivity analysis only concerns about whether the variable factors have had a huge impact.If the in?uence is big, we will consider,but for the study of probability is less. But in practice, people not only just looking for sensitivity factors, but also paying more attention to the product in its energy output, response latency, and other factors a?ect the happening probability,which guides the next work. Only these indicators reach the standard,the stimulation can successfully a?ect the individual. In project design, the ideal level is the experiment level that can successfully a?ect the individual with no less than probability P.Based on ideal level estimation for practical problems, this paper ?rst introduces former sensitivity analysis experiment method. Then, by applying sequential trial, this paper improves the probability P to meet the requirement. In Chapter 3, this new method is illustrated and proved. In Chapter 4, this paper veri?es the estimation e?ect of the new method by simulation.