Research On Travel Behavior And Travel Demand Forecasting Based On Household Decision

With the continuous development of Chinese economy, the travel demand of urban people has demonstrated an unprecedented quick increase. Travel demand management, on which has been put emphasis by researchers and administrators, is beginning to substitute the monotonous supply expansion. As a powerful implement of travel demand management policy, most of current travel activity analysis is based on individuals, which can not explain the many phenomena in activity decision and travel activity. According to theory of activities, travel decisions are made by households. Individuals'behavior will be affected and restricted by other members of the families. From a new perspective of household decisions, the author studied the theories, models, algorithms and applications of travel behavior and travel demand forecasting. In general, the study scope of this thesis includes the following contents.First, the history and current status of the research on three aspects, including the theories of human activities and behavior, activity-based travel behavior and travel demand forecasting, were introduced. The basic theory of household activity decisions was then illustrated and the impact of households on the travel behavior of their members was analyzed systematically. Based on that, the framework of household activity decision models was proposed.Second, due to its potential advantage in the area of travel behavior analysis, structural equation model was introduced into the area of activity analysis based on household decisions. The structure equation models for household activity analysis were then established and the fundamental theories, model structure and algorithms were studied.Third, based on the mathematical models and theoretical framework above, quantitative study regarding household activity pattern and travel behavior analysis in China was conducted using survey data. Based on the model result, both qualitative and quantitative explanations on the activity decision, activity assignment and trip generation of different type of households were given. Besides, the typical activity pattern of each type of households was also summarized.Finally, the travel demand forecasting models were established based on structural equation modeling. In addition, the author compared the models based on structural equation modeling with those based on neural network and discrete choice model, and demonstrated the models established by author were practical and advanced.Both the application and model comparison showed the models established in this thesis had great value in the areas of travel behavior analysis and travel demand forecasting.