Modeling the Dynamic Choice Behaviors in Ownership and Usage of Cars and Motorcycles
|關鍵字:||汽機車持有與使用;羅吉特模式;迴歸分析;馬可夫鏈模式;Ownership and usage of cars and motorcycles;Logit model;Regression model;Markov chain model|
Associated with the rapid growth of economic development in Taiwan, the ever-increasing number of private cars and motorcycles has inevitably brought severe problems of energy consumption, air pollution and traffic congestion. The problems become even more serious in urban areas. Thus, it is urgent and imperative to propose and implement management strategies which can effectively curtail the ownership and usage of cars and motorcycles for the goal of sustainability. In order to forecast the reduction in ownership and usage of vehicles under various strategies, to develop the choice behavior models of car and motorcycle ownership and usage is extremely important. Based on this, this study employs discrete/continuous Logit and regression models to model the car and motorcycle ownership and usage behaviors at national and regional levels, respectively. Then the proposed models are used to propose and analyze management strategies. According to the transitional probability determined by the proposed models, a Markov chain model is used to forecast the dynamic long-term growth rates of cars and motorcycles. To do this, this study first conducts a nationwide questionnaire survey by disseminating a total of 90,000 questionnaires to the owners of registered cars and motorcycles based on a stratified random sampling technique. A total of 5,986 valid questionnaires are returned. In addition, to further scrutinize the differences of choice behaviors in various living environments, 23 counties/cities of Taiwan are first classified into three regions: major metropolitan, minor metropolitan and ordinary city, then the regional car and motorcycle Logit models are developed accordingly. The calibrated results of the national ownership models show that social economics, living region and vehicle usage characteristics are major contributing factors to vehicle ownership. As to the regional models, fuel cost is significant in major metropolitan, but it is insignificant in other two regions, implying the gas price elasticity of dense-populated cities would be higher because of their convenient public transportation and congested traffic conditions. Although vehicle price is found to be significant in three regions, it has significantly different coefficients, which tend to be higher in a larger city. The reasons may be similar to that of fuel cost. As to the calibrated results of the national or regional usage regression models, the fuel cost is the most significant variable, indicating its major influence to the usage of vehicles. To investigate the applicability of the proposed models, three strategies, including the increase of gas price, the increase of vehicle price, and the fuel fee according to miles, are applied and examined. The results show that as the increase of gas price as 50% and 100%, the annual total traveling mileage of car will decrease by 23.44% and 41.39%, respectively; while that of motorcycle will decrease by only 6.10% and 11.80%. In case of the vehicle price increases by 50% and 100%, then the total traveling mileage of car will decrease by only 3.94% and 7.21%, respectively; while that of motorcycle will decrease by 1.35% and 2.35%. The policy of fuel fee according to miles will reduce the total traveling mileage of car by 5.22%.
|Appears in Collections:||Thesis|
Files in This Item:
If it is a zip file, please download the file and unzip it, then open index.html in a browser to view the full text content.