| As an important national infrastructure and one of the mass transport tools, the railway plays a decisive role in China’s economic and social development. It is the basic task in the railway construction of our country to develop high-speed railway, built rapid railway networks, which can meet the demand of large flow, high density, high speed and convenience. With the development of wireless communication technologies, train speed and traffic density have been further improved in moving block system. It is more prominently necessary and urgent to establish GSM-R wireless communication network with high confidence and the control system of dispatching optimization of train operation organization with high efficiency and low energy consumption. The network performance analysis of QoS in GSM-R wireless communication with high confidence and the characteristics of traffic flow of multi-train tracking in the communication with high confidence have been deeply researched in the thesis. The main research results and conclusions are as follows:1. The formal modeling of GSM-R wireless communication with high confidence has been considered and the high reliability has been analyzed under the condition of high-speed train running in moving block. The system model of wireless communication in CTCS described by stochastic Petri net and GSM-R fault recovery model between train and RBC have been established, and the corresponding conclusions have been obtained.2. For the problem of information parallel processing in wireless communication, a new membrane computing method based on biochemical reaction rate and stochastic process has been put forward. The biochemical reaction rate has also been achieved from the distribution function and assigned to the corresponding membrane. The quantitative analysis of information parallel processing and the reliability in wireless communication systems have been realized by Markov isomorphic transformation. The new method processes events in parallel manner, which has improved the efficiency of the system, and made up for the lack of practical application. Moreover, the method can improve the analysis accuracy of simulation system and provide a new way of the quantitative and qualitative analysis for the complex wireless communication.3. According to the characteristics of train running in moving block, the problem of train tracking on straight lines has been researched under the condition of wireless communication with high confidence. Based on the Multi-Agent theory, a MAS model of multi-train tracking has been established and the interaction mechanisms of MAS between train and RBC have been proposed, and the safe distance control of multi-train running has been realized. When trains are tracking, the effect of speed changes, initializing density of different lines on traffic has been simulated and the corresponding conclusions have been obtained.4. While multiple trains are tracking, the train flow will have real-time changes. It is difficult to statistically analyze the energy consumption for traditional methods, so a new modeling method of complex system based on snapshot layer has been proposed, which has been simulated on the energy consumption of train traffic. The time system can be taken as a snapshot at each time t, which includes static information and dynamic processor. With the evolution of a series of snapshots, the characteristics of the system are obtained. Combined with the established interaction mechanism of train tracking of MAS, the analysis model of energy consumption of train flow based on snapshot has been set up. The impacts of the initial density of lines and dwell time on traffic flow energy consumption are analyzed.5. The characteristics of train running flow on curve lines have been studied based on the NaSch model. A cellular automata model simulating multi trains tracking of railway curves has been proposed. The numerical simulation has been done by computer and the influence of different radius curves, outer rail superelevation and curve length on railway traffic flow has been studied by the model. |