Research On Travel Demand Forecasting Based On Complex System

With the rapid development of railway, "Market Demand Decided Transportation And Production" will become a railway transportation enterprise development guideline, travel demand forecasting highlights an important position, it is the prerequisite and basis of railway investment, train plan and pricing, medium and long-term travel demand forecasting will give a scientific basis for railway construction and upgrading the overall capacity, and short-term forecasting will provide effective support for transport capacity scheduling. Railway travel demand forecasting is no longer a fresh topic, as well as home and abroad experts,railway within and outside experts have studied many of models and algorithms, and continuous to improve to enhance their precision and to narrow the margin of error. However, travel demand is a system with high degree of uncertainty, which was very difficult to forecast a medium and long-term travel flow which just only uses an algorithm or model. It is very necessary to studies travel demand forecasting model from the perspective of complex systems which combined the complex characteristics of travel demand. From the perspective of complex systems, this paper analyzed the complexity of travel demand, the general travel demand forecasting model, the factors which affected travel demand and travel demand mode, and also pointed out the further studying issue.Firstly, this paper summarized the generally travel demand forecasting methods which evolved travel demand development mechanism and travel demand acts of space, based on this theory, gives a complex systems travel demand forecasting model (PFM-CS model) on the basis of the complex system, the model was described in detail, and pointed out the relations between the complex system travel demand forecasting model and the general travel demand forecasting model, and also pointed out the importance of the factors sets and parameters sets which was affected by the complex system travel demand forecasting model.Secondly, around the affecting factors of travel demand of the PFM-CS model, this paper made a detailed analysis from the time factors, space factors, travel demand properties factors, price factors and transport capacity factors and so all, the proposals was put forward for the value of the parameters analysis which was verified by actual travel demand data. Thirdly, this paper combined the thesis of PFM-CS model subsystem and vibration system which shared similar characteristics, and introduced the concept of mode to the railway travel demand forecasting model, the subsystem of PFM-CS model was saw as a one by one mode and analyzed of several commonly used method, in addition, it is more important to raise the mode processing steps and methods of mode division, mode choice, mode trends and dealing with seasonal component.Fourthly, around the PFM-CS model of the space levels, this paper used fuzzy clustering method (AGA-OFEM) to cluster analyze railway station, and divided into numbers of similarly level regional, and then select the Beijing area to apply PFM-CS model, the analysis of mode division, mode choice, mode trends and dealing with seasonal component was completed to verify PFM-CS model of science.Finally, combining the "Railway Passenger Transport Marketing Aided Decision System" of railway information technology planning, this paper demonstrated the methods and measures to complete the work and recommendations which applied the PFM-CS model to travel demand forecasting model, and also pointed out difficulties which need for further studies as mode model optimization, the travel demand affecting factors acquisition and quantify, mode division and mode choice.