频谱检测门限与n-out-of-K融合规则的联合优化

Joint Optimization of Spectrum Detection Threshold and the n-out-of-K Fusion Rule

为了减少计算复杂度,提出联合优化检测门限λ和n-out-of-K融合规则的算法。以λ和n为参数建立目标函数,并将参数以二进制形式表示,从而把算法转化为组合优化问题。接着,采用基于样值修改的互熵优化方法渐次逼近最优的参数。仿真表明,该算法在获得与已有算法几乎相当的总错误率情况下,可有效降低平均搜索次数,且随着K的增加搜索次数增加更平缓。

To reduce the computational complexity, an algorithm optimizing the spectnma detection threshold ）, and the n-out-of-K fusion rule jointly is proposed. A target function is built with the parameters λ and n. By binary encoding the parameters, this algorithm is converted to a combinatorial optimization problem. Then, the algorithm gains the near optimal results step by step by use of the modifying samples-based cross entropy method. Simulations show the proposed algorithm decreases the average search times effectively with almost the same total error rates as that of the existing one and the search times of the proposed algorithm increase more evenly than that of the existing one.