Study On The Fundamental Laws Of Riding Activities Of Urban Residents Based On Dockless Sharing Bicycles

Along with the rapid urbanization,we encounter many challenges,including environmental degradation,traffic congestion.To tackle various challenges induced by rapid urbanization,developing a greener and more sustainable transportation system is critical for the sustainable development of the future city,of which sharing transportation is an important component.With the great advances in Internet of Things(Io T)and mobile payment technology,sharing bicycles have experienced explosive growth in China.It greatly benefits residents’ daily life,yet,it also brings many practical problems.Previous literature mainly focuses on the study of the fundamental laws of human mobility,but lacks the analysis on the mobile characteristics of sharing transportation conveyances and the interaction between sharing conveyances and travelers.These two problems are of great significance for the design of more efficient and user-friendly bicycle-sharing systems,as in the future,with the further development of sharing economy and transportation technology,travelers would highly probably no longer possess their own transportation conveyances,but use sharing transportation conveyances.Currently,dockless bicycle-sharing systems can record the relevant information of each ride in detail,which provides a unique research opportunity to study the hidden patterns behind the riding behavior.Previous studies have found that human mobility of the whole travel mode obey some universal laws,for example,the distribution of travel distance and radius of gyration all follow a truncated power-law.In this paper,each bicycle is regarded as an individual entity and its mobility patterns is studied.After analyzing the riding data of tens of millions of sharing bicycles in six cities,it is found that the movement law of bicycles and users has no stable universal laws across different cities,and the mobility patterns of sharing bicycles is often very different from that of human beings.Although there are various significant differences in different cities,all the differences can be well explained by a universal power-law scaling behavior emerged from the interplay between users and bikes(i.e.,for users to determine which bike to ride).Such a scaling behavior shows that users will select the good bikes with a high probability,but they will also select not-so-good bikes with a not too low probability.The above findings can deepen our understanding of bicycle-sharing system and users’ choice behaviors,and contribute to designing a more effective system.For the bicycle-sharing system,an important practical problem is how to rebalance the distribution of bike supply.It is particularly critical for the sustainable operation of the system and the improvement of user experiences.By taking the spatial and temporal characteristics of bicycle flow networks into consideration,the accuracy of traffic prediction can be improved.Therefore,by integrating the regions and its relationship into graph structure,this paper proposes a prediction framework based on graph convolution network(GCN)framework.This framework integrates not only the spatial dependences between regions,but also the temporal patterns.At the same time,the prediction results of GCN are compared with several other classical regression models.The results show that GCN has a better performance than other models on predicting the flows between regions,which is very useful to the bike rebalancing of dockless bike-sharing systems and improve the experience of users when looking for bikes.The results of this paper also have broad and important applications in other similar scenarios in the era of sharing economy,and eventually assist inventing a greener,healthier,and more sustainable future city.