Study On Modeling And Optimization Of The Logistics System In Railway Container Terminal

The container transportation is a standardized, specialized, and internationalized modern transportation mode. As the critical node of container transportation network, railway container terminal is the bridge and tie between the railway transportation and other transportation modes. It plays a pivotal role in the container intermodal transportation, which means that its production capacity and operational efficiency has a huge impact on the entire container transportation network. Therefore, this paper focuses on the railway container terminal, researches and discusses the related issues of the actual production and management in order to provide the policy-makers and managers of railway container terminal with technical support and theoretical foundations of making plans and selecting schemes.In order to overcome disadvantages such as the singleness and one-sidedness of the individual prediction model, the combinational forecast model is applied to forecast railway container freight volume in order to raise the forecast precision. The forecast models are established by using polynomial curve prediction model and grey prediction model based on the historical data of freight volume. Then two individual prediction models are combined by the radial basis function neural network. The research results show that some error indicators of the combinational forecast model are the best compared with the two individual prediction models.Then, the combinational forecast results are analyzed and modified by Markov chain prediction model to increase the reliability.In consideration of the fuzziness of the container truck arrival rate, the in-out system of railway container terminal is a fuzzy queue system. The mathematical programming model which minimized the total system cost per unit time is established. Because the cost of waiting per unit time per container truck is fuzzy in the model as well, the concept of a-cut, the nonlinear programming and mixed integer programming method are applied. The lower and upper bounds of objective function are calculated by the different a levels. Finally, the membership function of the minimal total system cost per unit time is constructed, and the optimal service rate of in-out system is obtained in this fuzzy environment, which provides theoretical support for management and operation of the in-out system of railway container terminal. In addition, the situation with three fuzzy parameters is discussed. The above method can still work and get the corresponding analytical results, which expands the application scope of this method further.For the operation sequence scheduling problem of the front handling crane as the common device in the yard of railway container terminal, a linear integer programming model is established to minimize the total device operation time for all the tasks on the basis of integral consideration for the various constraints from the tasks and device. In order to solve this programming model, an improved bacteria foraging algorithm is proposed. A crossover operator, a hybrid replication strategy and a self-adaption migration probability based on the health degree and fitness degree are introduced during the processes of chemotaxis, reproduction, and elimination, respectively. These improvements can accelerate the convergence rate, restrain the escape behaviors of excellent individuals and the degradation phenomenon of solutions. The numerical example shows that the model and the algorithm are feasible and available.The internal logistics system of railway container terminal is composed of loading-unloading system, horizontal transportation system and stacking system, the optimal equipment configuration of which is researched based on the closed queuing network model. According to the operation contents and characteristics of the internal logistics system, a closed queuing network model of it is established. In this model, two service nodes of the rail mounted gantry cranes and front handling cranes as well as a virtual service node of transportation provide the logistics service for the internal container trucks. The approximate mean value analysis algorithm is applied to solve the closed queuing network model and calculate some equipment performance parameters, on the base of which the optimal equipment configuration is determined under the optimization conditions. Finally, a simulation model of the internal logistics system is established by the simulation software to examine the accuracy of the closed queuing network model and the approximate mean value analysis algorithm.With the view of the actual production and operation of railway container terminal, this paper pays close attention to the railway container terminal as the research object. The prediction method, queuing theory, fuzzy theory, optimization algorithm and modeling simulation are applied for the theory analysis and the method study of freight volume forecast, in-out system optimization, front handling crane scheduling and internal logistics system equipment configuration. The researches and explorations on the production and operation of railway container terminal have certain theoretical significance and important practical value, and lay a solid foundation for the future research.