中文摘要：

交通需求预测中会有很多不确定性因素．系统的不确定分析可以分析模型输出的可信程度，并且界定影响不确定性的关键因素．采用了基于敏感性分析的方法（SAM）分析交通需求组合模型（CTDM）的不确定性．CTDM基于出行者的理性行为，并可描述为一个非线性规划问题，从而分析敏感性．不确定性分析包括模型输入的不确定性和模型参数的不确定性．与基于随机抽样的分析方法相比，基于敏感性分析的方法在大大减少计算量的同时，可以得到近似的精度，并且可以同时或分开考虑输入的不确定性和参数的不确定性．研究认为，提高参数精度比提高输入精度能更有效地提高输出结果的可信水平．

英文摘要：

Travel demand forecasts are subject to great uncertainties. A systematic uncertainty analysis provides insight into the level of confidence in the model outputs, and can lead to identification of the critical sources of uncertainty. An uncertainty analysis is made of a combined travel demand model （CTDM） with the sensitivity-based analytical method （SAM）. The CTDM, on the basis of a single unifying rationale, is formulated as a nonlinear programming problem, and the sensitivity analysis is also made. The uncertainty analysis includes both input uncertainty and parameter uncertainty. The SAM, which requires less computational effort than sampling-based method in the uncertainty analysis, can obtain similar accuracy and can investigate the input uncertainty and parameter uncertainty separately. In order to improve credible level of outputs, to increase the accuracy of parameters is more effective than that of inputs.