City Air Quality Forecast Model And Software Development Research

Probability and statistics theory has broad application in research work today. The combine of the two theories has been the fundamental framework of any applying subjects, such as the theory of information, mathematics Risk, insurance for Actuaries and etc. And the forecast theorem, which derive from the historical data and deduce the statistics rule then get the future state of the system, is one of the best famous questions of these applying subjects.In 1978, Prof. Liuwen first applied an analytic technique in the study and got the theory --the strong law of large unmbers for transition probabilities of denumerable homogeneous Markov chains, which ground the theory of modern forecast application. The theory show its great applied importance in the field of environmental science and decision-making process. Especially in the application on air quality forecasting (AQF). The reason of fluctuant air quality is various, which result in multiplex factors. So the data always represent randomicity. In the last ten years, as the research of the AQF improving , Much means on quantitative or qualitative analysis have been made use of .And of all others, the most favorite way is statistical forecast theory for its economy,briefness and exactitude. Many models have been taken into application as its precision improving. This article based on Prof. Liuwen' theory and expanded its practicability .The author build a model of Markov chains for AQF , do some reach on the multi-analysis recursive models, comparability models, self-recursive models and then bring forward a new method which grounded on the two theories above for the design of operation software on AQF application.This article is established in doing reseach on AQF model grounded on the Markov chains theory. Much stress on the three aspects just as below:1. doing reach on the forecast model based on Markov chains with instance validating and veracity inspection. And draw a conclusion that the model is effectual and practical on the short-term forecast.2. doing reach on the forecast statistical model (multianalysis recursive models,comparability models,self-recursive models) .Applying instance validating and veracity inspection and analysing the veracity.3. after discriminating and comparing, achieve a new method grounded on the two theories above for the design of operation software on AQF application.