考虑决策过程与潜在异质性的居民通勤选择行为研究

Commuting Mode Choice Behavior Incorporating Decision-making Process and Latent Heterogeneity

传统研究对居民通勤决策过程的动态性及微观心理偏好关注不足,为此利用潜在类别模型表征居民通勤决策的潜在异质性,从动态决策角度出发,构建可以体现偏好动态演进过程的决策场理论模型。该模型对不同潜在类别设置差异化的属性注意力权重,表征不同子市场用户动态偏好的差异性和相同子市场用户偏好似而不同的特征,同时实现用户分类和动态决策过程分析,改进传统决策场理论对个体偏好异质性表征不足的缺陷。对成都市轨道交通车站附近居民进行实证研究,结果表明,通勤居民可以细分为“商务需求型”“追求效率型”“经济环保型”“保守型”这4类,各类别间的备选方案初始偏好、属性关注程度均呈现显著差异。通过Matlab仿真实现对个体决策结果、微观偏好进化过程及非理性效应的预测和分析,验证了本文方法的有效性。

Traditional methods pay little attention to the dynamics of the decision-making process and the micropsychological preferences of residents.This paper applies a latent class model to represent the potential heterogeneity in the commuting decision-making process.Then,from the perspective of dynamic decision-making,the decision field theory model is constructed,which can reflect the dynamic evolution of preference.The model sets different attribute attention weights for different subgroups,which can represent the heterogeneities of users'dynamic preferences in different subgroups and the similar but different characteristics in the same subgroup.Meanwhile,it can simultaneously implement user classification,analyze the decision-making process and improve the deficiency of traditional decision field theory in characterizing individual preference heterogeneity,based on which an empirical study of residents near rail transit stations in Chengdu is conducted.The results show that the residents can be subdivided into four types:business demand subgroup,efficiency demand subgroup,economy-preference subgroup,and conservative subgroup.There exist significant differences between subgroups in the initial preference and attention weights for attributes.The simulation realizes the prediction of decision-making results,the demonstration of the evolution process of individual micro-preferences,and the irrational effect in the process of attribute change,thus verifying the effectiveness of the proposed model.