Understanding variations in activity-based vehicle allocation decisions: A latent segmentation-based random parameter logit modeling approach

Abstract

This paper examines the activity-based vehicle allocation decisions among individuals during mandatory and discretionary activity-tours based on travel accompanying arrangements, specifically, solo travel (i.e. traveling alone) and joint travel (i.e. traveling with household/non-household members). Latent segmentation-based random parameter logit (LSRPL) models are developed in this study to explore the factors affecting vehicle allocation decisions within households, including travel characteristics, built environment and accessibility measures. For instance, model results suggest that presence of children in joint mandatory- and discretionary activity-tours increases individuals’ probability of getting SUVs from their households’ existing vehicle fleet. Also, tour complexity identified by higher number of activity stops, exhibits positive coefficient value for SUV allocation in case of discretionary activity-tours. One of the unique features of this study includes evaluating the effects of individuals’ attitudes on vehicle allocation decisions at activity-based tour-level. For example, due to positive attitude towards active transportation, likelihood of allocating any vehicle to the individuals decrease during both solo and joint mandatory activity-based tours. The vehicle allocation decisions in the households, however, vary across two segments. Older-higher income individuals in segment one tend to get SUVs from their household vehicle fleet during a joint discretionary activity-tour while living in higher mixed land-use areas, however, younger-lower income individuals in segment two exhibit a negative relationship for SUVs. In addition to the heterogeneity across segments, preference of SUVs during a joint discretionary activity-tour might vary among individuals within the same segment, as indicated by statistically significant standard deviation of ‘land-use index’ variable.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 7 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.