Abstract: For technical problem of the spectrum allocation for cognitive radio network, this paper presents spectrum assignment algorithm based Fitness Prediction Dual Artificial Bee Colony algorithm（FP-DABC）. The algorithm is designed to interfere threshold to improve the access number of users; fitness forecasting methods are used to speed up the assignment efficiency; at the same time the fairness of spectrum allocation process and system overall performance are optimized. Experiment results show: FP-DABC algorithm reduces a part of network bandwidth benefit, at the same time user satisfaction, allocation efficiency, the average assignment time, the user fairness and overall system performance are better than the Color Sensitive Graph Coloring algorithm（CSGC） and Artificial Bee Colony algorithm（ABC）.

Key words: fitness forecasting, artificial bee colony algorithm, cognitive radio, spectrum allocation, allocation ratio

摘要： 针对认知无线网络频谱分配过程存在的问题，提出了基于适应值预测策略的双人工蜂群算法（FP-DABC）。该算法设计的干扰门限阈值，提高了用户的接入数量；适应值预测方法的使用，加快了分配效率；同时算法对频谱分配过程公平性和系统整体性能进行了优化。实验仿真结果表明：FP-DABC算法牺牲了部分网络效益的同时，在用户满意度、分配率、平均分配时间、用户公平性和系统整体性能上均优于颜色敏感图着色算法（CSGC）和人工蜂群算法（ABC）。

关键词: 适应值预测, 人工蜂群算法, 认知无线电, 频谱分配, 分配率