Intelligence Computing And Its Application In Network Optimization

Optimization consists in many fields and it also has huge application future during the country economy development. The issues combinational optimization studies relate to a good many fields such as information technology, economy management, industry project and traffic transportation, such as network optimization. Because of the non-linearity, restrict nature, multi-target, multi-mode, even non continuum or non parse of the target function, it is difficult to solve the combinational optimization problems by traditional numerical value method which have based on the strict mechanism model.Compared with traditional optimization method, the advantages such as non-central-control, multi-agent mechanisms, simple structure, inner parallel and easy to be understood and realized, effectively promote the development of intelligence computing in optimization technology. Intelligence computing play more and more important roles in optimization of production process, improving the production efficiency and benefit and saving resource. The intelligence computing methods such as Evolution Computing and Swarm Intelligence, by means of imitating life-form and bionic simulation, adapt to the computing system which have the intelligence character of large scale parallel, auto-organization, auto-adaptation and self-study, and can offer fruitful method and approach to solve some complex problems.The main idea of this paper is to research the theory of Evolution Computing and Swarm Intelligence Computing, the algorithm innovation and its representative application to network optimization. In subject selection, it starts with Evolution computing and Swarm Intelligence computing. It makes the Genetic Algorithm (GA) as its study foundation. It makes the novel Artificial Fish Swarm Algorithm (AFSA) and the combination of DNA Algorithm and Evolution Algorithm as its cut-in point. In theory, it presents systemic expatiation and study about the principle, structure and implement methods of the algorithms told above. In application, it applied the improved algorithm to the typical network optimization problems such as routing optimization of the computer network and the coverage optimization of WSN which is the hotspot and difficulty of the WSN research. In algorithm innovation, it syncretizes the technique of taboo table of Taboo Search Algorithm into AFSA as well as adds a parameter. In DNA-GA algorithm, it solves the difficulty of coding and also syncretizes the technique of DNA algorithm into the design of gene-pass-operator and the aberrance operator. In the effect of simulation, the two improved algorithms build up the ability of global optimization and neighborhood search and get better convergence rapidity and optimization efficiency. In research method, it pays attention to the landscape orientation analysis and summarizes the parameter characteristic of the three algorithms above and of the evolution computing and swarm intelligence computing, further more, it emphasizes the improvement and development direction of the algorithms.The paper starts with the concept, system and main characteristic of the optimization problems and Intelligence Computing, and introduces the Combinational Optimization, the Evolution Computing, the Swarm Intelligence Computing and the network optimization. Then, the GA is studied and applied to the routing optimization of computer network optimization. The simulation results and analysis are given. The third, the principle, convergence capability and implement method of AFSA are expatiated, the improved algorithm is presented, the simulation results, analysis compared with GA and improvement way are given. The forth, the structure, process and implement of DNA-GA is studied. The new algorithm was designed according to the coverage optimization of WSN, the simulation results, analysis compared with GA and improvement way is given. Finally, the paper gives the parameter characteristic of the three algorithms above and the characteristic of the Evolution Computing and Swarm Intelligence Computing, further more, the development of the algorithms.