Dr Sisi Jian

Research Associate & Postdoctoral Teaching Assistant
School of Civil & Environmental Engineering

UNSW Sydney

Dr Sisi Jian is a Research Associate at the Research Centre for Integrated Transport Innovation (rCITI), School of Civil and Environmental Engineering at the University of New South Wales (UNSW Sydney) in Australia. She received her PhD degree in Transport Network Planning from UNSW Sydney, master's degree in Logistics from Nanyang Technological University, and bachelor's degree from Central South University. Her major research interests include shared-mobility (carsharing, ridesharing, bikesharing shared autonomous vehicle), big data in transport (application of social media data and crowdsourced data in transport research), machine learning, demand and behavioural modelling, and network modelling and optimisation.

Understanding and Optimising Carsharing Systems

One dominant challenge in carsharing systems is to ensure the supply of vehicles can meet the demand of users in a cost-effective manner. This requires accurately predicting users' demand and optimally relocating vehicles in response to demand variations. The two principal areas of this thesis are methods to estimate demand and optimally relocate fleet.

 

From the demand side, this study models users' vehicle selection and utilisation patterns. A spatial hazard-based model is proposed to investigate users' vehicle selection behaviour. Upon making a vehicle selection, users then decide the amounts of consumptions to allocate to each selected vehicle type. This process is modelled by the multiple discrete-continuous extreme value model. These two models are calibrated using data from an Australian carsharing company. The findings can help operators determine the most efficient allocation of resources.

 

From the operation side, this research develops and solves novel models for the vehicle stock imbalance problem in one-way carsharing systems. Two models are proposed to link demand and supply, both incorporating a discrete choice model (DCM) in a Mixed Integer Programming (MIP) model. The MIP model solves optimal relocations and updates vehicle availability; the DCM coupled with the updated vehicle availability changes users' demand reciprocally. The results reveal if there is a strong interdependence between demand and supply, the supply has a critical impact on system profit.

 

The core contribution of the research is to take the first attempt to understand and optimise carsharing systems considering the interdependency of demand and supply comprehensively.