A novel excess commuting framework: Considering commuting efficiency and equity simultaneously

Abstract

Excess commuting, which concerns the differences between the actual commute and the optimal (minimum) commute afforded by a given distribution of jobs and housing, i.e., urban form, has been extensively studied across disciplines. In the existing excess commuting framework, the optimal commute considers commuting efficiency but overlooks commuting equity, which is defined as the variation in commuting cost across workers before and after the optimisation. The framework also overlooks the variation in commuting frequency across workers for a period of interest, which also affects the overall commuting cost for the period. In this paper, we propose a novel excess commuting framework using a Greedy-Initialisation-based Genetic Algorithm, where the optimal commute accounts for commuting efficiency and equity and commuting-frequency variation simultaneously. We illustrate and calibrate the framework using one-month metro smart card data in Shanghai. Comparing with two other existing models, the Greedy-Initialisation-based Genetic Algorithm can generate a commuting pattern that balances commuting efficiency and commuting equity, which the existing commuting framework and corresponding algorithms cannot.

You are how you travel: A multi-task learning framework for Geodemographic inference using transit smart card data

“…Sari proposed a heuristic model to detection home locations from London's Oyster card data, but it was only applicable for tube users because the alighting bus stops were missing in SC data. Additionally, in previous works, the home locations detected from SC data are mainly used for demand estimation or commuting/travel behaviour analysis (Zhang, Zhang and Zhou, 2019). Hence, the results of home location identification are always aggregated in the bus/tube stations.…”