A Study On Four-Step Travel Demand Model Incorporating Induced Traffic

New or improved road is one method to mitigate traffic congestion, which can increase road capacity. However, new road construction and improvement will generate new travel demand (named induced traffic in this paper) in certain time. Therefore, the induced traffic by road new construction or improvement needs to be taken into consideration completely during the traffic demand forecasting. It is necessary to formulate the prospective road transportation planning in order to meet future road travel demand.This paper aims to establish a four-step travel demand model incorporating induced traffic. The main contents include three aspects as follows.Firstly, study on the definition, mechanism and elasticity models of induced traffic. Based on the current situation of domestic and foreign studies on induced traffic, we clearly define the notion and mechanism of induced traffic. The elasticity model method, which is an important research tool in the induced traffic study, is applied to analyze the current situation of induced traffic in China, based on the factors which may influence vehicle kilometers of travel (population, Gross Regional Product and lane-kilometers et al) in the following body of researches. Several basic elasticity models are included, among them:the elasticity-based model, delay model, distributed lag model, autocorrelation model, autoregressive distributed lag model, growth model, dummy model and fixed-effect model; also discussed are advanced elasticity models, such as the three stages of least squares. Results show that lane-kilometers in China are found to have a statistically significant relationship with vehicle kilometers of travel measurements of about0.026-0.274in the short-term (1year) and0.367-0.773in the long-term (more than2years, including2years). The detail of result analysis confirms the existence of induced traffic in certain time after road improvement. Meanwhile, the induced traffic in the short-term is generated gently, and in the long-term increased significantly.Secondly, considering the accessibility in traditional four-step travel demand model, four-step travel demand model incorporating induced traffic is constructed with feedback. In such model, two shortcomings could be overcome. In the first place, the feedback construction will solve the problem of inconsistent travel time in traditional four-step travel demand model. In the second place, the accessibility in four-step travel demand model incorporating induced traffic can reflect the influence of improved accessibility on further travel demand after new road construction or improvement. This will solve the problem of without consideration induced traffic in traditional four-step travel demand model.Finally, based on the four-step travel demand model incorporating induced traffic with feedback construction and parameter estimation (including road traffic impedance function), such model is tested in TransCAD software. Dalian city is selected as a case study using related data collected in both1994and2010, getting three mian results. In the first place, compared results of this model and traditional four-step model certify that four-step travel demand model incorporating induced traffic can predict more accurate traffic volume in reality. In the second place, following the procedure of four-step travel demand model incorporating induced traffic on TransCAD, the generated induced traffic in Dalian city from1994to2010accounts for about9.7%of all traffic in2010. The elasticity between lane-kilometers and vehicle kilometers of travel turns out to be0.372, suggesting that a0.372%growth in vehicle kilometers of travel happens when lane-kilometers increase by1%. In the third place, based on the road network in2010, using the four-step travel demand model incorporating induced traffic predict the travel demand in2020. This shows the Dalian’s traffic distribution in the future.In conclusion, the proposed four-step travel demand model incorporating induced traffic has two advantages. It can not only predict the induced traffic, which is caused by increased road network accessibility after road constructions and improvements, but also improve the prediction accuracy of traditional four-step travel demand model. These both can provide precise traffic forecasting in urban transportation planning with future travel demand.