CONTACT US: Contact info
|
Citation |
Cai Y, Ong GP, Meng Q. Transp. Res. C Emerg. Technol. 2023; 147: e103990. |
|
Copyright |
(Copyright © 2023, Elsevier Publishing) |
|
DOI |
|
PMID |
unavailable |
|
Abstract |
Bicycle sharing systems (BSSs) are expanding with unparalleled speed all over the world due to various advantages of cycling such as cost effectiveness, environmentally friendliness, capability to bridge other public transport modes and reduce private vehicle usage. One of the latest trend of BSS mode appears to be free-floating, which offers more flexibility in bike rental and return for cyclists in terms of service locations. However, despite the convenience brought by the free-floating BSS to the users, it also causes a headache for the authorities, largely attributed to its poor enforcement on regulating bad user behaviours, which significantly deteriorates the city scape. In order to preserve the benefits of BSSs and meanwhile solve the problems caused by free-floating BSS, this study develops a multi-objective integer non-linear programming (MOINLP) model to convert an existing free-floating BSS into a geo-fence-based one by determining the geo-fencing stations to be opened, their corresponding parking capacities and deployment of bikes. The three objectives of the developed MOINLP model are minimizing monetary cost, user dissatisfaction as well additional first-/last-mile walking distance of users. The MOINLP model is transformed into an equivalent multi-objective integer linear programming (MOILP) model. The augmented ε-constraint method (AUGMECON2) is customized to solve the MOILP model for exact Pareto optimal solutions while a tailored simulation-based heuristic method is proposed to solve large-scale problems. The numerical experiments have testified the good performance of our proposed solution method and we extend our analysis to Beijing BSS to gain planning insights and policy implications. Language: en |
|
Keywords |
Augmented -constraint method; Bicycle sharing system; Fleet size; Geo-fence; Location capacity; Multi-objective optimization; Simulation-based heuristic method |