摘要
认知无线电的性能优化是一个动态多目标优化问题。现有的Bio-CR模型基于遗传算法优化认知无线电的性能,它使用线性加权方法将此多目标优化问题简化为了一个单目标优化问题。针对Bio-CR很难确定每个适应度函数的权值和容易漏掉一些最优解的问题,提出了基于多目标遗传算法的认知无线电性能优化算法CREA。CREA能够根据信道条件和用户服务需求的变化动态地调整传输参数以优化性能,不仅克服了Bio-CR的两个缺点,而且通过保存计算结果进一步减少了遗传算法的运行次数。CREA首先根据信道条件的变化动态确定一组适应度函数,然后运行多目标遗传算法获得一个Pareto-optimalset,最后根据用户服务需求从中选出一个最满意解,并通知认知无线电更新自己的传输参数。Matlab仿真实验证明了CREA的正确性和有效性。
The performance optimization of cognitive radio is a muhiobjective optimization" problem.It's difficult to assign the weight of each objective,and some optimal solutions will be omitted while using linear weight method to simply the multiobjective optimization problem into a single object optimization problem.We develop a new performance optimization algorithm called CREA based on a Muhiobjective Evolutionary Algorithm.At first, CREA decides a group of fitness functions according to the changing of channel condition,and then runs the Muhiobjective Evolutionary Algorithm to get a Pareto-optimal set.At last,CREA chooses a most satisfying solution from the Pareto-optimal set according to the users' service requirement and notifies the radio to update the transmit parameters.The results of simulation experiment testify the validity of CREA.
出处
《计算机工程与应用》
CSCD
北大核心
2007年第20期159-162,共4页
Computer Engineering and Applications
基金
国家自然科学基金(the National Natural Science Foundation of China under Grant No.60602019)