Abstract:

In order to optimize the structure of travel mode choice forecasting model, and deal with the limitations of utility theory in terms of individual risk appetite, incomplete rational decision making, and overall utility of travel mode, a two-factor travel plan is constructed by combining travel mode with departure time, and a joint model based on Nested Logit-cumulative Prospect Theory is established. The objective utility and selection probability of the NL model are subjected by cumulative prospect theory, and two functions (cumulative weight function and value function) are constructed to describe the actual perceived value of travel mode to travelers in the form of prospect value. Finally, the forecasting model is calibrated and tested by the survey data. Results show that the prediction accuracy of the Nested Logit-cumulative Prospect Theory model is higher than NL model, and comprehensive hit ratio rises from 74.8% to 85.2%. The prediction hit ratio of each mode is more balanced.

Key words: urban traffic, travel mode choice, cumulative prospect theory, Nested Logit model, perceived value

摘要：

为改进效用理论在个体风险偏好、非完全理性决策、方式整体效用等方面的表述局限，联立出行方式与出发时段构建双因素出行方案，并建立基于巢式 Logit(NL)-累计前景理论的出行方式选择预测优化模型.通过累计前景理论将 NL模型所获方案客观效用及选择概率主观化，构建累计权重函数、价值函数并以前景值的形式描述出行方式对出行者的实际感知价值，最后通过调查数据进行建模与验证.结果表明，与仅基于 NL模型进行的方式预测相比，所建模型综合命中率从74.8%上升至85.2%，各方式预测命中率更为均衡.

关键词: 城市交通, 出行方式选择, 累计前景理论, Nested Logit模型, 感知价值