Satisfactory Optimization Theory And Its Application In IP Network Optimization

In recent years, with rapid development and permeation of IP networks, the construction of IP networks can be seen in everywhere of the world. In this situation, the study of networks optimization is becoming more and more important,which can help construct economical, high performance and stable networks.The traditional optimization design focuses on getting the "best" solution, but because of most networks optimization problems usually belong to multi-constraint and multi-objective optimization problems, the traditional optimization design are NP-complete problems. To obtain the "best" solution needs not only complex algorithm, but also a long time during the algorithm runs. The time can become so long that the algorithm cannot be used to solve the practice problems. In this dissertation, a concept of satisfactory solution is proposed to take place optimum solution, a theory of satisfactory optimization is proposed to study networks optimization problems.This dissertation analyzes the limitation of the optimum theory, reviews the study of satisfactory theory, studies satisfactory multi-objective optimization deeply and systematically, and establishes the satifactory multi-objective optimization model. This paper uses this model to be the evaluation system of optimization scheme, and selects GA(Genetic Algorithm) to be the search method searching the potential optimization scheme. The evaluation system and GA are unified the same architecture to become a complete solving model of satifactory multi-objective optimization. This dissertation uses the solution model to study the backbone network optimization and the QoS (Quality of Service) multi-cast routing.In the term of design a computer communication network, especially a backbone network, it needs to select a route for each communicating pair and a capacity value for each link in the network to get minimum overall cost of the network. In addition, the utilization of the whole network needs also be considered. The proposed solution model is used to solve the backbone networks optimization problem.In the same time, a suitable GA is designed to solve this backbone network optimization problem. According to the computing of realexamples, the solving model is proved that it can gain high quality solution, in the same time, it convergent fastly.As a result of the increasing demand of multimedia applications, efficient and effective support of quality of service (QoS) has become more and more essential. QoS routing is one of the key technologies of realization of grarantee QoS applications. Multi-cast is an important style of QoS applications, which consists of concurrently sending the same information from a source to a subset of all possible destinations in a computer network. Multi-cast utilizes a tree as delivery structure, on which data packets are duplicated only at fork nodes and are forwarded only once over each link. The multi-cast routing algorithm focuses on construct a good performance multi-cast tree, on which every kind of application's QoS is ensured.This dissertation analyses the principle of multi-cast and multi-cast routing technology, uses the satisfaction multi-objective optimization solution model to construct a QoS multi-cast tree. A suitable GA is designed to search the potential optimization schemes. According to the simulation on the random produced networks, the solution model proposed in this dissertation shows a good convergence speed and good ability of loading balance of network stream, compared to the already proposed algorithms of establishing a multi-cast tree such as KPP, BSMA, CDKS.