改进型自适应随机卷积网络编码算法

Improved Adaptive Random Convolutional Network Coding Algorithm

To address the issue of field size in ran-dom network coding, we propose an Improved A-daptive Random Convolutional Network Coding (IARCNC) algorithm to considerably reduce the a-mount of occupied memory. The operation of IARCNC is similar to that of Adaptive Random Convolutional Network Coding (ARCNC), with the coefficients of local encoding kernels chosen uni-formly at random over a small finite field. The difference is that the length of the local encoding kernels at the nodes used by IARCNC is constrained by the depth; meanwhile, increases until all the re-lated sink nodes can be decoded. This restriction can make the code length distribution more reasona-ble. Therefore, IARCNC retains the advantages of ARCNC, such as a small decoding delay and partial adaptation to an unknown topology without an early estimation of the field size. In addition, it has its own advantage, that is, a higher reduction in memo-ry use. The simulation and the example show the ef-fectiveness of the proposed algorithm.