Abstract: Genetic Algorithm(GA) and adaptive GA often trap into local optimum when they are used in optimizing the system utility of Cognitive Radio(CR) network. By introducing the Euclidean distance reflecting characteristics of genes, this paper improves GA, and proposes a spectrum allocation scheme with system utility maximization. Experimental results prove that compared with traditional GA and adaptive GA, the scheme proposed has better performance.

Key words: Cognitive Radio(CR), spectrum allocation, Genetic Algorithm(GA), Euclidean distance, system utility maximization

摘要： 传统遗传算法和自适应遗传算法在优化认知无线电网络的系统效益时易陷入局部最优。针对该问题，通过引入反映每个染色体基因特性的欧氏距离，对遗传算法进行改进，提出一种系统效益最大化的频谱分配方案。实验结果表明，该方案的频谱分配性能优于基于传统遗传算法和自适应遗传算法的方案。

关键词: 认知无线电, 频谱分配, 遗传算法, 欧氏距离, 系统效益最大化

CLC Number: 

• TN925