Prediction Methods Of Key Indexes On Road Transportation Planning

The development of national economic has made for the steady improve of transportation, and it also has brought up higher standard for traffic service level. Government departments and experts had been committed to Intelligent Transport System.. Road traffic is part of Intelligent Transport System. The intelligence level of road traffic is related to regional economic development and improvement level of people's living. The accurate prediction of road transport planning index is its focus question. So exploring the prediction methods in road transport planning key indexes not only provides rationale but also support theoretical basis for traffic planning.The scientific road transport planning can promote developments of society and economy. Road transport planning indexes includes demographic indexes, economic indexes, transport indexes and public transport indexes. The passenger volume, cargo, passenger turnover, freight turnover and bus passenger volume play a key role in transport planning. The transport indexes provide basis for passenger and freight network and layout of passenger and freight station. Bus passenger volume provides basis for the bus real-time vehicle dispatching, vehicle optimization and allocation.This paper study the relevant indicators of road traffic planning, it mainly study the forecasting method of loss of traffic, passenger and freight transport and public transportation passenger's flow. The foreign and domestic experts had developed different forecasting methods, which had provided richness theory and valuable experiences for sequential studies. But the theory and methods still have some defect, it should be intensively studied. Less prediction model studied the inherent laws of objects and affect factors, but complexity within the system can result the large error, which made the models accurately forecast an object of study. Combination forecasting models are insufficient to combine the features, and it is difficult to achieve effective integration. so the prediction results are always not acceptable. The study of public transportation passenger flow is less. It changes with time and the data distribution is nonlinear. So existing models can hardly predict its tendency.This paper continues to conduct in-depth study based on previous research, explore the method of solving problems, construct the prediction model fitting the road transportation planning index and enrich its prediction method theory system. The main contents include the following:1. Improved system dynamics prediction method:The paper analyzes the characteristics of the system dynamics, it studies the inherent laws of objects and influence factors with the angle of system. Based on the relation of system variables, the paper constructs System dynamics model. Considering System complexity resulting large error, the paper has improved the model and developed improved system dynamics prediction method.2. Gray neural network ensemble forecasting model:The paper analyzes the characteristics of the Gray model and neural network model. It applies GM(1,1) to construct 1-AGO sequence. Gray simulation sequence is produced by gray model and network model. Then the paper applies it to construct gray neural network ensemble forecasting model.3. Prediction method based on morlet wavelet neural network:The paper analyzes the characteristics of morlet and applies it to network model. It adds scaling factor and translation factor to model neuron and draws the model structure. Then prediction method based on mother wavelet neural network is constructed.4. The wavelet decomposition and reconstruction neural network prediction method:The wavelet decomposition and reconstruction method is introduced to neural network model based on its characteristics. The passenger flow data as signals are decomposed into approximation and detail signals by wavelet filter. The approximation and detail signals are predicted and reconstructed for forecasting passenger flow volume. The wavelet decomposition and reconstruction neural network prediction method are proposed finally. The method is used to forecast the short-bus traffic.5. Empirical analysis:The proposed road transportation planning prediction methods are applied to practices. For the research scope of Jilin province, passenger transport amount, passenger turnovers, freight transport amount and freight turnovers are predicted in practice, and the predicted results are compared with 'Outline of Jilin Province Road Transportation Twelfth Five-Year Plan. The predicted errors are analyzed to verify the effectiveness of prediction methods.Innovation of this paper in the following areas:1. For overcoming the large errors caused by system complexity. Estimation error module is introduced to the model for effectively reducing forecast error effectively.2. Grey neural network model is constructed for overcoming the shortcomings of grey model and neural network. And also the model overcome the defect of demanding lots of input data. The method is used to forecast road traffic in the case of small sample size.3. Morlet neural network model is constructed and the scaling factor and translation factor are introduced to the model in order to extract information features effectively. The model can avoid the local optimization problem by adjusting parameters. The application of the model improve the accuracy effectively. The method is used to forecast road traffic in the case of no significant variation of the data.4. The chaos theory is applied to the prediction of passenger volume. The paper reason out that traffic passenger flow data is chaotic for the first time. The wavelet decomposition and reconstruction is introduced to neural network model based on its characters. With complex variation of passenger volume, the research solve the problem that traffic passenger flow is hard to predict. Test results shows that the method has higher prediction accuracy.The research results of this paper will develop and enrich prediction methods theoretical framework of traffic planning index. The research results will provide basis for drawing up road transportation development planning and road traffic specific plan.