A Hierarchy of Deep Reinforcement Learning Agents for Decision Making in Blockchain Nodes

【Author】 Abu Mallouh, Arafat; Abuzaghleh, Omar; Qawaqneh, Zakariya

【Source】IEEE EUROCON 2021 - 19TH INTERNATIONAL CONFERENCE ON SMART TECHNOLOGIES

【影响因子】

【Abstract】Distributed P2P network applications are increasingly growing and it is expected to gain more attention in the future for the wide range of attractive features it offers. One of the popular applications of P2P networks is the blockchain technology and associated applications such as cryptocurrencies. The blockchain technology was developed to operate on a distributed environment of nodes without the need for a central authority. One of the challenging tasks facing applications that operate on P2P networks is the collective decision-making process between the distributed nodes for reaching a consensus between the peer nodes. In this paper, a framework is proposed to handle the problem of collective decision making between distributed peer nodes. The framework introduces the usage of distributed intelligent agents which will be trained by deep reinforcement learning. Three models have been developed and evaluated using randomly generated data. The best achieved result by the proposed framework was obtained by using Model 3, the overall accuracy of the model was 82.2%.

【Keywords】Deep Reinforcement Learning; Blockchain; Neural Networks

【发表时间】2021

【收录时间】2021-12-31

【文献类型】会议

【主题类别】

区块链技术--