Demand Management of Station-Based Car Sharing System Based on Deep Learning Forecasting

Abstract

Metropolitan development has motivated car sharing into an attractive type of car leasing with the help of information technologies. In this paper, we propose a new approach based on deep learning techniques to assess the operation of a station-based car sharing system. First, we analyse the pick-up and drop-off operations of the station-based car sharing system, capturing the operational features of car sharing service and the behaviours of vehicle use from a temporal perspective. Then, we introduced an analytical system to detect the system operation concerning the spontaneous deviations derived from user demands from service provisions. We employed Long Short-Term Memory (LSTM) structure to forecast short-term future vehicle uses. An experimental case based on real-world data is reported to demonstrate the effectiveness of this approach. The results prove that the proposed structure generates high-quality predictions and the operation status derived from user demands.

Understanding User Behavior in Car Sharing Services Through The Lens of Mobility

“…The reason may be that the car sharing usage during a long-term period is evolving, which results in the deep learning-based methods challenging to capture the dynamics without user behavior consideration or overfitting, while XGBoost-based methods normally show excellent performance for the problems with small-to-medium structured/tabular data, which is exactly our extracted user behavior related features. The difference between our XGBoost and BeXGBoost is that XGBoost does not consider the user demographic features and latent revisitation patterns, which are rarely considered by existing bike sharing and car sharing usage prediction works [34,54,57,58]. Another reason may be that the advantage of the user behavior highly-related features we obtained from our qualitative and quantitative user studies, which potentially indicates the significance of our work on user studies.…”

“…We compare our user behavior-aware BeXGBoost with the following state-of-the-art prediction methods, which are widely adopted in recent bike sharing or car sharing usage prediction [54,57,58,60].…”

“…We found users have different purposes for using car sharing systems, which result in different spatial usage patterns. Hence, we divide the city into seven categories of functional areas based on the method in [59] (i.e., residence, entertainment, business, industry, education, scenery spot, and suburb) to capture the spatial characteristics of service stations ℱ the usage distance, i.e, the average usage distance of each station Convolutional Network (GCN), which are widely used in bike sharing and car sharing usage prediction recently [34,58,60].…”

Vehicle Assignment considering Battery Endurance for Electric Vehicle Carsharing Systems

“…is requires system to be long-term reservation type, or strongly assume demand can be accurately and stably predicted [35,36]. Up to now, many studies are exploring demand prediction [37,38] but still not quite accurate and robust enough to drive the optimization. Towards the instant access type, the system presents strong randomness and dynamics; simulation approaches perform well to model the system characteristics.…”

Recurrent Neural-Based Vehicle Demand Forecasting and Relocation Optimization for Car-Sharing System: A Real Use Case in Thailand

“…e second category is the operator-based strategy, in which the vehicle operator forecasts the demands at each station and then sends the operation staff to relocate the vehicles according to the demand forecasting results. is strategy ensures that the vehicle relocation tasks are done as needed but requires highly experienced staff to work for the vehicle relocation [8,[12][13][14][15][16][17][18]. In [12,15], solutions based on the mathematical models and rule-based algorithms were proposed.…”