A digital decision approach for scheduling process planning of shared bikes under Internet of Things environment

Digital technology plays an important role in the construction of intelligent transportation systems. The digitization of travel and traffic information contributes to the efficiency, equality, and safety of mobility for urban residents. This research aims at improving the imbalance between the supply and demand of bike-sharing system. A scheduling process planning algorithm for shared bikes towards Internet of Things (IoT) environment was proposed in this research. Firstly, the algorithm combined the massive shared bikes data of IoT, Long Short-Term Memory and Gated Recurrent Unit neural networks (LSTM-GRU) hybrid model was employed to predict the shared bike demand towards electronic fence(e-fence). Then, based on the division of scheduling sub-areas, the mathematical model of scheduling path optimization for shared bikes was constructed, in addition to the constraints and costs, the minimum carbon emission was also considered as the objective function. Finally, the algorithm was applied to the bike-sharing system in Yanqing District (Beijing China) as the case study. The results show that: the LSTM-GRU hybrid model proposed in this research has a high prediction accuracy of 6.16 mean square error and 0.86 goodness of fit, the sub-area partition model can reduce the average scheduling demand imbalance degree by 58.90%. Under the same scheduling task, electric vehicles can reduce carbon emissions by about 32% compared with fuel vehicles. The proposed shared bike scheduling planning algorithm can provide a decision-making guidance for related operation departments, as well as realize low-carbon and sustainable development for urban transportation system.