Robust Stage Plan And Plan Adjustment For Railway Marshaling Station

Marshaling staions are the primary units of the railway network, and the key nodes of the rail hub and the main rail lines. The mainly works of marshaling stations are handling the classification and assembly of freight trains, and make sure the trains are departed on time according to the departure schedule. Those yards operations are all planed by yards dispatcher. There are two steps of yards plans in China, and the most important one is the stage plan. There are many uncertainties when formulate and carry out the stage plan. While few researches focus on this topics that the optimization of the stage plan under uncertainties. This dissertation studied the robustnees stage plan and plan adjustment for railway marshaling stations. The main works of this dissertation are as follows:1. According to research trend of the stage plan for railway marshaling stations, the current situation as well as the necessity of the research on robust stage plan are reviewed and depicted. The review are based on five aspects, such as wagon-flow allocation, yard engines utilization, yard track utilization, and the stage plan optimization under certain and uncertain environment.2. Analysised the theory of optimizating the stage plan of railway marshaling station. The necessity of the research on robust stage plan is depicted after the analysis of the uncertain of the plan information. The concepts, characteristic and evaluation factors are proposed. Based on them, the framework for robust stage plan is put forward, that is the framework of "plan formulation+plan adjustment" for marshaling station.3. Robust wagon-flow allocation (WFA) models are studied. Three wagon-flow allocation models are proposed step by step. Firstly, a WFA model that considers different train size limitations of departure trains is formulated. The weight and length of departure trains are taken into account as the size limitation simultaneously. And the cars number and the freight value or benfit of departure trains are considered together for a more general WFA model. Secondly, a WFA model considers the uncertain of the breakup time of inbound trains and the makeup time of outbound trains is formulated. We estimate those time information more accurately by taking the cuts of trains, the classification process and assembly process into account. So the uncertain impacts on the stage plan can be reduced if the gaps of estimated and actual values are decreased. Thirdly, an equity objective function is introduced for the WFA model that considers the uncertain of arrival time of inbound trains. This equity objective function can optimiza the distribution of waiting time for classification of the inbound trains that from different directions.4. Based on the robust WFA model, the robust stage plan optimization model is presented by adding the study of yard track utilization problem. Then the model is decomposed into three sub-models:yard engines utilization model, static WFA model, and yard track utilization model. And different algorithms are proposed for all models. For example, the harmony search for the yard engines utilization model and ILOG for the yard track utilization model. Then a numerical case is used to verify the proposed mathematical models and the solution algorithm. The robustness of the stage plan is analysised through the redundancy of cars, the gap between estimate and actual classification time and assembly time, and the waiting time of inbound trains for classification and so on.5. The theory of strategy adjustment for stage plan is researched. Based on the framework of "plan formulation+plan adjustment", this dissertation proposed the theoretical and methodological framework of strategy adjustment for stage plan. This framework including the definitions, forms, conditional branches, and the definition of strategy meta and so on. The algorithms proposed in the previous sections are improved for adapting to the disturbance of the stage plan informations under the dynamic environment. The knowledge and experience of dispatcher must be highlighted during the adjustment of stage plan. For example, the strategy meta can be present by dispatcher, and the strategy must confirmed by dispatcher befor adopted. At last, some numerical case is used to verify the strategy adjustment framework and solution algorithm.