Forecast of Time Variant Parking Demand by Time Sequence Models for Multi-Type Hotels in Urban Centers

摘要

Focused on the parking spaces of leisure and convention hotels in urban centers, this research is conducted in the following sequence: a comparison in the fluctuation characteristics of parking demand between weekend days and work days is made in the beginning. Afterwards, the demand forecasting models in time sequence method are established. Then, a case study of the parking demand forecasting on a workday for convention hotels is applied. As the evaluation indicators as well as error analysis indicate, an ARTMA (3,1,0) model is rather appropriate for the prediction of the short term parking demand of convention hotels on a workday. As the results reveal, the first difference sequence of the parking demand for convention hotels is confirmed as stationary series, while an original stationary sequence is for leisure hotels, whether it be a workday or a weekend day. The parking demand of leisure hotels possesses lower amplitude fluctuation and therefore is relatively more stable.