Validation Methods And Assistant Tools Based On Coherence Of Data

Model is the base of simulation. According with users'purpose, the quality of a model will seriously affect the judgment of the real world or the confidence of using a result from a simulation. Model VV&A is a valid approach of ensuring model quality and model reliability evaluating. Validating is an important part of VV&A,where data coherence analysis is a common method in.Surrounding the data coherence analysis, research was carried through.Firstly, there are many data analysis methods, which using range is different from each other. By far, a whole data analysis system for different instance does not exist. But, a system like this will improve the efficiency of model validation and the precision of validation's result. According to the goal that cover most instances in data analysis, an analysis system is brought forward in this article,Secondly, model validation not only ensures model quality, but also provides evidence for reliability evaluating. Model validation always works with a index system which bottom node adopts data coherence analysis. The index system is also used in reliability evaluating. With bottom node's fusing, we will get the value of root, which denotes the coherence between model and real world and is an important part of reliability. In this article, it is called 'accord degree". But bottom nodes adopt different analysis methods. So the results are not in the same space and we can't fuse them directly. In this article, five methods are given to solve this problem.Finally, data coherence analysis always needs complex calculate. It requires that validation people have abundance knowledge about analysis method. To improve efficiency and precision of data analysis, an assistant software is programmed in this article using VC6.0.