Research On Service Path Optimization For Container Trains Based On Sea-rail Intermodal Transport

"One belt,one road" strategy has made the demand for opening up and international trade in all parts of China,especially in the inland areas,have increased sharply.However,due to the unreasonable transportation structure,the economic and environmental benefits of import and export logistics in inland areas can’t be effectively developed.Network stability is also difficult to guarantee.In view of the advantages of large railway transportation volume and complete transportation network,railway transportation can be used to replace some highway transportation to complete the inland collection and distribution of the port,so as to solve the problems of traffic congestion and environmental pollution.Moreover,in the process of long-distance transportation,railway transportation is easy to produce a certain scale of economic benefits,which can effectively reduce the transportation cost.Therefore,the establishment of a sustainable sea rail intermodal service network has become one of the important measures to rationalize the transportation structure.The collection and distribution efficiency of ports has always been the bottleneck restricting the development of marine transportation.This paper mainly studies the container collection and distribution operation.From the perspective of sea rail intermodal transportation,this paper proposes to use efficient railway container trains to complete the assembly and distribution in inland areas.In the supply environment with uncertain demand,how to accurately predict the future container throughput and reasonably arrange the port operation,so as to avoid the stacking cost caused by cargo accumulation in the port or the time cost caused by untimely supply,is one of the urgent problems to be solved in the collection and distribution connection operation in the port operation area.The close combination of accurate prediction of port container throughput and optimization of train service path can comprehensively improve the timeliness of port container collection and distribution operation,meet the needs of customers for timeliness of transportation,save costs as much as possible and improve the efficiency of "door-todoor" transportation of the port.Based on the above background,this paper studies container throughput prediction and train route optimization.Aiming at the complexity of port container throughput data,a short-term hybrid prediction model of container throughput based on optimal variational mode decomposition and kernel extreme learning machine was proposed.Firstly,the outliers was removed by Hampel Identifier(HI)from the original data,the preprocessed time series was decomposed into several sub modes with obvious characteristics by the Optimal Variational Mode Decomposition(OVMD).Then,to improve the prediction efficiency,the decomposed sub modes were divided into three categories by the values of Sample Entropy(SE): high frequency low amplitude,medium frequency medium amplitude and low frequency high amplitude.With the help of Kernel Extreme Learning Machine(KELM)involving the different prediction performance kernel functions,wavelet,Gaussian and linear kernel functions were used to capture the trend of sub modes emerging different characteristic,respectively.Finally,the final prediction result was obtained by linearly summing the prediction results of all sub modes.The proposed model has advantages in prediction accuracy and efficiency.At the same time,it overcomes the defects such as modal aliasing in traditional Complementary Ensemble Empirical Mode Decomposition(CEEMD)and Ensemble Empirical Mode Decomposition(EEMD)and over fitting in Extreme Learning Machine(ELM),and has practical application potential.Since uncertain factors are affecting the operation of container trains in the process of searail intermodal transportation.Combined with the customers’.demand for a fixed time window,the uncertain planning interval is introduced to represent the range of time in container loading and unloading at each customer node.Meanwhile,the demand time window with timeliness requirements is set as a soft constraint.The penalty function is integrated into the objective function of the transportation cost as a penalty term.A reasonable penalty coefficient is selected to construct a multi-objective optimization model of the train service path combined with the low transportation cost and less transportation time.For uncertain variables,the chance-constrained programming transformation model is used to obtain a Multi-objective path optimization model considering fuzzy time.Then,the multi-objective problem is transformed into a single objective problem by weighted summation,and the artificial bee colony algorithm is designed to solve the constructed model.Taking Shenzhen port sea rail intermodal transport as an example,the model is tested and compared.It shows that the model and algorithm can well meet the needs of different customers with different transportation timeliness,and has obvious advantages in transportation cost.