Study On Data-driven Vessel Scheduling Model And Method For Restricted One-way Inland Waterway Transportation

Due to the low cost,less pollution and large freight volume of the inland waterway transportation.The state council of China issued "Guidelines on Yangtze River Golden Waterway and Economic Belt Promotion" and "Yangtze River Economic Belt Comprehensive Three-dimensional Transport Corridor Plan(2014-2020)" which explicitly expressed the plan of improving the traffic capacity、strengthening the shipping ability and establishing the intelligent service system and security system.The traffic capacity of inland restricted waterways is very low which has become a bottleneck for the inland waterway development,the study on improving traffic capacity in inland restricted waterways is of great theoretical and practical significance.Scheduling the ships to pass through the restricted waterway in a reasonable sequence is the most practical way to improve the traffic capacity in the restricted waterways,However,this not only requires accurate prediction of ships’ driving time and passing time but also needs to face the uncertainties in the restricted waterway traffic.Firstly,this paper reveals the relationship of ship speed and the waterway circumstances by studying the speed variations of ships passing through the restricted waterways.After that,this paper establishes a ship trajectory length prediction model by studying the overall trajectory and analyzing its influencing factors.And then design a ship scheduling and sequencing strategy to guide ships passing through the restricted waterways quickly.Finally,propose a ship speed guidance based scheduling and sequencing method to further reduce the waiting time of ships.The main works of this paper are as follow:(1)Take the speed of ships passing through the restricted waterways as research objects,propose a new neural network construction and optimization method to model the long-term speed when ships passing through the restricted waterways.The method gradually increases the complexity of the neural network until the relationship of input and output data can be fitted.To obtain the parameters of the neural network,intelligent optimization methods and the orthogonal least square algorithm is adopted to ensure the orthogonal relationship of hidden neurons for a compact neural network.(2)Research on the ships’ trajectories in the inland restricted waterway areas,propose an overall trajectory length prediction model construction method.Dividing the overall trajectories into serval segments to highlight the trajectories’ local characters for data preprocessing.Obtain the similarities between trajectories based on unsupervised method.Analysis the latent relationship between ships’ characteristics and ships’ trajectories.Then calculate the trajectory length by probability theory.(3)Propose a real-time ship scheduling and sequencing method base on the built ship long-term speed model and overall trajectory length prediction model to guide ships pass through the restricted waterways quickly.This method decreases the waiting time of ships and improves the traffic efficiency by arranging an optimized sequence to ships while passing through the restricted waterways.(4)For the problem of upstream ships usually spend too long time to wait for the traffic signal,propose a speed guidance based ship scheduling and sequencing method to further reduce the waiting time of upstream ships while ensuring the priority of downstream ships.The experiment results indicate that this method can enhance the experience of upstream ships while travelling through restricted areas and reduce the total waiting time,the traffic capacity is thus further improved.