基于Q-Learning的无线区块链广播路由算法

Wireless blockchain broadcast routing algorithm based on Q-Learning

面向无线区块链网络场景,参考常用的区块链逻辑分层,提出一个独立灵活的分层逻辑架构,详细地介绍了每层的工作原理及作用。在现有广播路由算法的基础上结合无线区块链网络场景,针对广播过程中存在的广播冗余及广播总能耗高等缺陷,提出了一种基于Q-Learning的广播路由算法。该算法将无线区块链网络交易信息广播过程描述为多个agent的广播决策问题,通过线下学习、训练,帮助每一个当前广播节点agent做出是否广播的决策。仿真实验结果证明,该算法有效地解决了广播冗余问题,减少了广播过程的总能耗。

An independent and flexible layered logic architecture for wireless blockchain network scenarios is proposed with reference to the commonly used blockchain logic layering,and the working principle and role of each layer are introduced in detail.Based on the existing broadcast routing algorithm and combined with the wireless blockchain network scenario,a Q-Learning based broadcast routing algorithm is proposed to overcome the defects of broadcast redundancy and large total energy consumption of broadcast in the broadcast process.The algorithm describes the wireless blockchain network transaction information broadcasting process as a broadcasting decision problem for multiple agents,and helps each current broadcasting node agent to make a decision whether to broadcast or not through offline learning and training.The simulation experimental results prove that the algorithm effectively solves the broadcast redundancy problem and saves the total energy consumption of the broadcast process.