Research On Allocation Algorithm Of Bicycle-sharing Parking Spots Based On Multi-objective Genetic Algorithm

Bicycle-sharing gradually replaced the public bicycle to solve the public transportation "the last kilometer" problem,to a certain extent changed the way people travel,but the bike rental location management is too loose,the vehicle mobility rush hour cluster question is easy to cause traffic congestion,has serious impact on the cycling resource management.Therefore,it is particularly important to improve the allocation of resources to Bicycle-sharing.At present,the main mitigation methods are the planning of stop points and cycle dispatching management.However,users' parking awareness is not perfect,and the demand for human resources in cycle dispatching is too high to fundamentally solve the problem of stop point congestion.With the launch of the moped industry,enterprises began to restrict the rental location of bicycles,and the way to guide users to park gradually appeared in the public's view.Based on multi-objective genetic algorithm,this paper guides the stop point allocation of Shared bikes,mainly including the following contents:To solve the problem of vehicle location information privacy and excessive search space of nearby stop points,geohash algorithm is adopted to convert user coordinate information into one-dimensional string,change the format of storage data,reduce the search space of nearby stop points,and protect user privacy.Aiming at the contradiction between shortening bicycle parking distance and dispersing bicycle gathering,a distribution model based on multi-objective genetic algorithm is proposed,in which the distribution sequence of vehicles is taken as the gene sequence of chromosomes,and regression algorithm is used to replace mutation operator in the descendant generation step of genetic algorithm,and the algorithm is improved.The experiment used Hypervolume evaluation index,and the results showed that the genetic algorithm integrated with regression thought could obtain a higher Hypervolume fraction,indicating that the population has better convergence and distribution.After the regression algorithm expands the data set continuously,the execution time of the original genetic algorithm will be affected,the regression operation will be asynchronous to the genetic algorithm,and distributed computing will be adopted.Although the rate of new individuals generated by the regression algorithm will gradually decrease,the execution time of the algorithm will be successfully reduced.Figure[23]table[6]reference[70]...