The Computer Network Routing Research Based On Granularity Computation Intelligence

The computational intelligence is an inspiration from the nature (including animal nature)'s principle in order to imitate, design and solve problems. It is the theory and method which takes the models (including computational model and mathematical model) as the foundation, uses the distribute-parallel computing as the characteristics, and simulates person's intelligence to solve problems.At different levels the observation, understanding, representation of the real world problems and the analysis, synthesis, reasoning process are obvious feature of human problem solving abilities. In some sense, this is just the intelligence in human problem solving process. In reference to the ability and feature of human problem solving, the artificial intelligence researchers did in-depth research on it, and established various abstract models. Two academicians, Professor Zhang Bo and Professor Zhang Ling proposed the problem solving theory based on quotient granularity space from the viewpoint of bionics. Their theory conforms to the thinking pattern of human problem solving. For such a complex system of network routing, there will be many more promising topics such as using granularity space problem solving theory from different levels and different angles, unifying the qualitative thinking and quantitative analysis organically, and analyzing and disposing the issues of routing strategy.Routing strategy is one of the core technologies of computer networks. The rapid development of Internet makes the routing strategy an important research direction of computer science. Research of the routing strategy becomes the key theoretical topic of the new generation network architecture and agreement theory. This dissertation summarizes the results of the previous studies in this direction, and makes improvement on the existing routing strategy and its realized algorithm. Based on the granularity theory the author analyzes the OSPF routing strategy. Furthermore he develops some new studies on computer network routing strategy and proposes some new algorithms and models.In this dissertation the study is based on the computation intelligence of quotient granularity space and its application on computer network routing strategy. The main contents include:1. It introduces the theoretical method and application of granularity computation intelligence. It discusses the reasoning model of quotient granularity space theory and synthesizes theory of domain, and analyzes the application significance of quotient granularity space theory in the field of computer network routing. Then it applies the quotient space theory to OSPF routing algorithm, proposes the computation intelligence method based on quotient space, which is the application and realization at actual network routing research.2.On the comparative analysis of network load balancing and resource optimization for routing algorithm, the routing algorithm based on link flow assignment weight and the OSPF link weights optimization based on self-adapting genetic algorithm, the author establishes the cluster,zone,group granularity model for computer network routing based on quotient space granularity theory. The dissertation describes the dynamic computer network routing algorithm based on granularity intelligence and carries the algorithm simulation in the OSPF network environment. The results manifest that the granularity routing theory greatly reduces the granularity search space of problem solving, and lowers the computing complexity in practical application. So this result has a significant guidance for enhancing the traditional OSPF routing efficiency.3. The dissertation introduces the thinking method and some new widespread application method of Multi-agent system in recent years at information technology domain. According to the multi-agent system as a very complicated system, this dissertation proposes the new method that Multi-agent can be applied to large complex dynamic network routing based on quotient space granularity theory. It establishes the quotient space granularity model and granularity agent theory. So the agent can be divided into cluster,zone,group agent by granularity using the inherited nature of no scaling network the hierarchy relation of unceasingly decrease. It not only solves the uncertainty of belief with time variation which is difficult to be hold, but also lowers the complexity of the system. It significantly reduces the amount of computation. Applied to the field of complex dynamic network routing strategy, it greatly improves the efficiency of network routing. So the author provides new ideas and methods for problem solving of routing. The empirical test results show that the theoretical model and algorithm which this thesis proposes are suitable for solving large dynamic network routing problem. The model and algorithm have some improvement on the convergence response and the efficiency of network routing system.The most effective means of promoting scientific progress is application. Quotient space theory has widespread application in the problem solving domain by its solid theoretical basis and rather perfect mathematical model. This research is only a start on the computation intelligence based on quotient space and its application at routing strategy. In the field of computation intelligence research the quotient granularity space theory combined with other theories will be a direction which deserves in-depth research. Along with the in-depth study of computation intelligence method based on quotient space, the author will gradually apply new methods to the field of computer network routing algorithm. The further study will contribute to the research and development of computer network routing.