Summary |
As the awareness of sustainable development rising, how to promote green transportation
become a vital issues to the government. Public Bicycle System (PBS) is one of the most
popular projects in many cities globally. However, a cautious long-term decision could lead to
a successful PBS since the infrastructure with high capital investment is unlikely to change
within a short period. This study focuses on the public station location problem as an operating
planner by building a two-stage stochastic programming model. In this model, the fixed cost
of setting a station, the parking cost per public bicycle and revenue per unit of demand are
given, whilst the most profitable location set and the capacity of each station will be determined
in the first stage. After considering demand uncertainty, price rate is the re-course decision in
the second stage. Additionally, due to the fluctuate demand in PBS affected by a variety of
factors, this study develops a demand function includes two main factors, access distance and
usage fee. In order to solve this non-linear mathematical model, this study develops a solution
method based on Genetics Algorithm which is used widely in location problem and gradient
descent method which provide an efficient way to search the optimal price rate and the capacity
of the stations. A set of numerical experiments and practical-scale problems are performed to
validate the developed model and solution algorithm. |