摘要
引入信息量偏差概念,给出平均交互信息量关于输入概率的增量公式,设计出离散信道容量线性乘法迭代和线性常系数迭代算法,它们都优于现有的指数迭代算法.并证明在所有单步迭代算法中它们几乎是最好的算法.
Introducing an notion of information-capacity-distance, this paper gives a increment formula of average alternate information capacity on input probability and designs a linearitymultiplied-alternate and a linearity-coefficient-alternate algorithm for computing the capac- ity of arbitrary discrete memoryless channels. These algorithms do better than exponentalternate-algorithm in existence, and they are proved almost the best algorithm in all single step alternate algorithms.
出处
《应用数学与计算数学学报》
2006年第2期19-27,共9页
Communication on Applied Mathematics and Computation
基金
浙江省教育厅基金资助项目(编号:20040507)
关键词
离散信道容量
信息量偏差
迭代算法
capacity of discrete channels, information-capacity-difference, alternate algorithms