Grey Modeling Technique And Its Application On Road Traffic Accidents Management

Grey system theory is a system science to solve problems of “poor information and small amount ofsample data”. Grey modeling technique is the core of Grey system theory, and also the bond of GreyTheory and its application. Grey modeling technique has successfully solved a large number ofpractical problems that can’t be solved by traditional methods. And there is a wide applicationbackground and strong application value of grey modeling technique. This dissertation is aimed at greymodeling technique and its application on road traffic accidents management, hoping to work out somenew research results of grey data transformations technique, grey correlation analysis, grey predictiontechnique and practical application. It can provide new ideas, new methods for grey modelingtechnique, expand the range of application, enrich the grey modeling system, and also promote theintegration of grey system theory and traffic management. The main research contents of dissertationand research results are as follows:(1) By analyzing the mechanism of the grey data transformation technique to the improvement ofsequence modeling, clear the structure mechanism of grey modeling technique. As the grey datatransformation technique can increase the smoothness and narrow the level deviation, propose theconstruction principles and theoretical guidance for the grey data transformation. Construct weakeningbuffer operator with variable weights based on the average growth rate. This operator can not onlysatisfied the construction principles of buffer operator, but also be able to adjust its intensity bychanging the weights, enriching the buffer operation system. Construct function-type datatransformation by linear function and anti-cotangent function, and verify its feasibility, expanding thefunction-type data transformation system.(2)Considered the changing characteristic of periodic sequence, propose a method to determine theperiod and how to measure volatility of the periodicity sequence. Considered the periodicity andvolatility as factors of grey correlation analysis model, construct the grey correlation analysis modelwhich meets periodic sequence. It could be a model support to analyze the correlation between periodicsequences. According to the characteristics of panel data, grey correlation analysis model based onpanel data is established, it’s a new idea to analysis the correlation matrix sequence. Propose a methodto determine the timing weight based on new information priority principle, and a method to determinethe index weight based on discrepancy actuating principle. Then, construct a dynamic multi-indicatorsdecision making model based on the timing weight and index weight.(3) Since the traditional GM (1,1) model is difficult to predict the volatility sequence, firstly deal with the volatility sequence by amplitude compression transformation and weighted mean valuetransformation, in ordor to weak the random fluctuation and improve the smoothness of sequence,and flatten the gradient of sequence. Through the transformed data to establish GM (1,1) model, newideas for the prediction of wave sequences provide new ideas. According to the system evolution stagecharacteristics, which are an embryonic phase, uniform change phase, accelerated phase andequilibrium, GM (1,1, t)model is constructed to reflect the dynamic evolution of system. Then makea research on the modeling mechanism, model characteristics, parameter estimation and other issues.Considered the characteristics of traffic behavior variable, parameters adjustable GM(1,1) power modelis constructed. Determine the range of parameter by grey information covering theory, work outthe optimal parameter by particle swarm intelligence algorithms search, and make intelligentoptimization of GM (1,1) power model. Combine GM (1,1) model with grey correlation analysis,reduce the possibility that non-key factors are elected to GM(1,N) model by grey correlation analysis.(4) Based on the research on grey data transformations technique, grey correlation analyses, andgrey prediction technique, analyze the causes and evolution trends of road traffic accidents in China bygrey modeling technique, and get the main impact factors of road traffic safety. Then, make aprediction analysis on the evolution trends of road traffic accidents in China. According to the causesanalysis and the prediction analysis of road traffic accidents, targeted measures to enhance road trafficsafety are proposed.