Structure Design Methodology For The Railway Intelligent Transportation Systems

Optimal design methodology for the structure of Railway Intelligent Transportation Systems (RITS) is developed in this dissertation. Recent development in distributed networking technologies has set more new challenges in the system design process. To settle the uncertainty and complexity in designing a distributed complicated system, current state of methodology arts needs to expand for the system development content in fields of structure design and system optimization. As a large distributed information system, RITS involves modern control, real-time scheduling, business operation management and such theoretical fields to provide different services for physical distributed customers or designers. To feed these multi-service and multi-task requests, methodology of how to design the system structure plays a most critical role in the development of RITS. Subsystems of RITS must fit into a dynamic framework to adapt continuous adjustment of requests. The methodology of organization of RITS surely decides the final outcome. With suitable new system design theory, RITS developers would introduce good technology performance and optimized economic purpose into such a large complex system. RITS can come out with fine-grained components for interoperable dynamic constructing, while with improving system reliability, robustness, reusability, and maintenance.Based on the research works entitled "Study on Architecture of the Railway Intelligent Transportation Systems", this thesis presents and formalizes, for the first time, a physical structure optimization design for RITS, and a set of optimized indexes with respect to methodology are presented. This dissertation does some research on RITS melioration design and structure optimization by using systems engineering, artificial intelligence, and computer networks, etc. Three important issues have been introduced by this research. Firstly, Splitting a complex RITS into different fine-grained components by using fuzzy clustering technique. Secondly, it focuses on how to optimize the physical structure design process of RITS by considering technique factors and economic factors with various configurations of physical elements under certain resource constraints. Thirdly, finding the design method for system dynamic reconfiguration, which is for stabilizing the functions of big system when sub-system sinks. To sum up, processing the system optimize design by adopting the methods which show above, they keep RITS physically effective, economical, reliable, and extensible. Based on theory above, the results achieved in this thesis have been applied to an emergency rescue information system showing the effectiveness in mapping a logic structure of a RITS to its physical structure. The major contributions of this thesis include:l.It brings forward the design ideas of optimize mapping by layers from the logic structure of a RITS to its physical structure. To resolve the problem about the complex physical structure design, on the base of the decomposition by the different size sub-system, these ideas depress the complexity of system design; bring logic structure design and physical structure design together simultaneity.2.A function cell polymeric method, which is based on fuzzy clustering, has been presented. At the same time, two concepts, "meta-service", "critical path " are introduced by this dissertation and it presents a method for "meta-service" extraction. With the help of system engineering, fuzzy mathematics, a RITS logic structure model based on a fuzzy association degree matrix is built to analyze the association degree in quantity between system processes, and fuzzy clustering methods are explored to classify logic functions of a RITS. In term of granularity principle of strong cohesion and loose coupling, it partitions the system into many independent function cells, which lay a foundation for optimization mapping of RITS from logic structure to physical structure.3.A comprehensive and practical performance evaluation infrastructure for RITSs has been presented, which can easily evaluate system performance and cost. Evaluation criteria are defined for calculation in quantity.4.To break the limitations of traditional RITS design, which is based on software engineering. A formal description of physical structure of a RITS is given for the first time. Multi-objective based optimization methods are used to meet design goals utilizing system resources in an effective way, which means low cost, rapid response, high reliability and good collectivity performance.5.To establish a multi-objective optimized RITS physical design method, which can resolve the RITS structure design problem, such as multi-variable, multi-object, multi-restriction. It presents two types of optimization methods-Generative Method and favored method for physical structure optimization design of RITS. It is hard to design a common framework for optimizing RITS by using traditional optimization methods due to the complexity, i.e., its optimized-index function and constraints. The first type of methods is a generative Pareto GA method; the research key point is multi-objective GA. Using modified individual sorting methods, virtual fitness assignment strategies, survival methods, a multi-objective genetic algorithm ~ NSGA, is improved to obtain aoptimization sets for physical structure of a RITS. The generative Pareto can effectively reduce the complexity of calculation and thus speeds up the process of convergence. The second type of methods, Preference Pareto GA, combines multi-objective optimization for physical structure design of a RITS with multi-objective decision making methods to present an expert system based two system design methods. One is based on fuzzy preference, and the other is based on fuzzy logic optimization. One main feature of the two methods is that they both integrate the preference information or experience from decision makers in the optimization design process to have an efficient system design. The optimization methods above have been applied to an emergency rescue system for physical structure optimization design and satisfactory results have been achieved.6.An application layer, based design approach for RITS has been presented, which integrates fault tolerance and disturbance-free mechanism in a consistent way. It also presents the basic principle of redundancy configuration of system nodes and the structure of RITS fault tolerance system, on the base of theory of information protocol, it brings forward a dynamic data stream reconfiguration algorithm, which is based on distance-vector algorithm. A better fault tolerant algorithm has been conceived using fuzzy logic based evaluation methods. In the mean time, system redundancy also makes it possible to execute jobs in parallel and increases the throughput of the system in total.