基于改进引力搜索算法的认知无线网络频谱分配

为了提高认知无线网络频谱资源利用率，提出一种基于改进引力搜索算法的频谱分配方案。该方案 将质子位置信息映射为频谱分配变量，提出一种非线性引力衰减因子，以增强算法对于频谱分配可行解的搜索 能力。其次，为了保持算法对于频谱分配可行解的最优搜索方向，采用全局最优引导策略为质子增加记忆能力。 然后，提出动态高斯扰动策略，随着算法迭代自适应地调整扰动步长，避免算法陷入局部最优。最后，针对频 谱分配问题的离散化特性，引入Sigmoid 函数使质子的位置离散化。仿真实验结果表明，改进引力搜索算法能 够有效地实现频谱分配问题的优化求解，取得了相对较高的网络效益和用户公平度。

Spectrum Allocation Based on Improved Gravitational Search Algorithmin Cognitive Radio Networks

To improve the utilization of spectrum resources in cognitive wireless networks, a spectrum allocation scheme based on improved gravitational search algorithm is proposed. In this scheme, the location information of the agent is mapped into spectrum allocation variable, and the nonlinear attenuation factor is proposed to enhance searching capability of the algorithm for candidate solutions of spectrum allocation. And then, to maintain the optimal search direction of the algorithm for the candidate solutions, the global optimum guided strategy is utilized to add memory function to the agents. Moreover, the dynamic perturbation operation based on Gaussian distribution is proposed to avoid the local optimum, which is adjusted adaptively with the increasing of iterations. Finally, the Sigmoid function is introduced to discretize the position of agents for the discretization of spectrum allocation. Simulation results show that the proposed algorithm achieves better network utility and user fairness than the comparison algorithms.