DIRAC: A Dynamic Programming Approach to Rateless Coded Multi-Hop Multi-Relay Transmission

Xijun Wang,Wei Chen,Zhigang Cao,Min Sheng,Jiandong Li
DOI: https://doi.org/10.1109/pimrc.2013.6666521
2013-01-01
Abstract:Owing to the capability of accumulating mutual information from the transmission of previous nodes, rateless codes can boost the network performance considerably, and hence have sparked much interest recently. However, how to efficiently schedule multi-hop multi-relay transmissions with the aid of rateless codes remains a challenging work. Particularly, it requires high complexity to find an optimal route due to its inherent combinatorial nature. In this paper, we formulate the optimal transmission scheduling as a dynamic programming (DP) problem by defining a novel state and constructing a tree-structured state transition diagram. It is from a point of view of DP that we further propose two low-complexity algorithms, namely S-DIRAC and Fano-DIRAC, with negligible performance loss based on the idea of sequential decoding of convolutional codes. Simulation results indicate that the low-complexity algorithms almost always find the optimal solution and show the superiority of routing with mutual information accumulation compared to conventional shortest path routing.
What problem does this paper attempt to address?