A Multiple Population Genetic Algorithm Based On Population Cluster

Genetic Algorithm is an effective global optimization search algorithm, and it is widely used in many artificial intelligence fields. It has the characteristics of simple, parallel with strong robustness. This paper makes a study about the Genetic Algorithm, proposing Double Population Chaos Genetic Algorithm and Multiple Population Genetic Algorithm Based on Population Cluster, basing on the Double Population Genetic Algorithm.The research background and the search mechanism are introduced in paper, and considering about the disadvantage of easily maturity and easily falling into local optimization, we propose the Double Population Chaos Genetic Algorithm. It seems that the chaos system is stochastic and out of order, but it has its own exquisite structure. We can use it's characteristic of stochastic, ergodicity and sensitivity of initial value to initialize the population, making the individuals evenly distributed, avoiding the defects of slow convergence caused by the totally stochastic. With the chaos disturbance, the algorithm can jump out of the local optimization, increasing the ability of getting the optimal solution. Through the tests by classical examples, the results show that this algorithm is excellent at the aspects of average truncated generation and on line performance.Based on the Double Population Chaos Genetic Algorithm, this paper proposes the concept of Population Cluster and cluster head, summarizing the Genetic Algorithm Based on the Population Cluster. Each Population Cluster has three populations with different evolution pattern, and the first focus on the local search, the second focus on the overall search, while the third that is the cluster head can balance the search ability at both the local search and overall search through the migration pattern. Through the tests by classical examples, the results show that this algorithm makes progress at the aspects of average truncated generation and on line performance. At last, we use this algorithm in the application of Cognitive Radio, and receive good affect.