Consider The Research On Bike-Sharing Scheduling With Self-Adjustment Between Stations

In recent years,with the rapid development of the sharing economy,the bike-sharing industry has played an important role in easing urban traffic congestion and building a green transportation system.However,under the macro-control of the bike-sharing industry by the government,the bike-sharing industry has entered the "quota era".For enterprises,under the new quota mechanism,no matter how difficult it is for the traditional operation idea of "grabbing the site" to work,the survival of the weak will be faster.How to respond to the assessment indicators to solve the problems of excessive output and high dispatching cost has always been a hot research topic in the current enterprise operation and management process.Based on the analysis and research at home and abroad on the basis of previous work,according to the new confronted by Shared cycling scheduling problem,namely the added ability between sites,path repair vehicle recycling needs,and there are many unreasonable bicycle supply three core problems,put forward from the demand forecast to scheduling path optimization of one-piece solution,and through comparing with previous work proves the validity of this method.The details are as follows:(1)A bike-sharing demand prediction model based on pseudo-double-hidden layer feed-forward neural network is proposed.From the aspect of network model structure,PDLFNs not only considers the advantage of multi-hidden network structure to improve the prediction accuracy of the model,but also considers the generalization ability of single hidden layer.From the aspect of sample feature extraction,PDLFNs not only fully considers the periodic factors of the time series model,but also carries out multi-dimensional mining on the factors affecting the demand for bicycles.(2)A dynamic scheduling model of Shared bikes considering self-adjustment between stations is constructed.Model fully considers the new problems encountered in the actual scene in scheduling,site of the bike according to the scheduling requirements set three types of tags: call in,call up,maintenance,and then consider car capacity,scheduling time window,simulate the speed of live,models,etc.To maximize customer satisfaction,and minimum operation cost and the shortest total time scheduling objective function is established.Finally in the aspect of algorithm,this paper puts forward two kinds of similar group of intelligent algorithm,can effectively avoid algorithm respectively,based on the mathematical deduction and problems of the improved particle swarm algorithm,and by updating the rear-end behavior and improved artificial fish algorithm of adaptive adjustment of step,in a Shared bicycle borrowed car validated data sets,and the final scheduling scheme is given.(3)The strongly feasible solution and the weakly feasible solution theorems are proposed,and a targeted two-stage operator is designed.On the basis of the work in the third and fourth chapters,in order to effectively deduce the logical relationship between the variables in the material allocation problem model with the characteristics of self-adjustment,the formulation of the problem parameters is firstly derived,the concepts of strong feasible solution and weak feasible solution are defined,and the lemma and proof are carried out.Then through formula deduction design fit into the class characteristics of two phase operator scheduling problem itself,and finally with heuristic algorithm,in a Shared hcho bicycle borrowed car data sets,and the other from the perspective of intelligent algorithm to solve the scheduling problem,compared in this paper,from the problem of extracting operator method with heuristic algorithm to solve the problem with fast cutting solution space,and search for the characteristics of the effective feasible region,can greatly improve optimization performance of the algorithm.