Integration of blockchains in wireless networks poses scalability issues. This is due to the fact that mining delay exponentially increases with number of blocks, thus causing bottlenecks during packet transmissions. This research paper aims to overcome the scalability issue, by proposing a novel reinforcement learning Deep Dyna Q model for improving the Quality of Service (QoS) in Blockchain-based Internet of Things (IoT) networks. The proposed model is designed to efficiently learn and adapt to dynamic changes in network conditions and provide optimal QoS for IoT devices & deployment sets. Deep Dyna Q combines the benefits of deep neural networks and Dyna architecture to effectively model the selective-encryption blockchain and improve the decision-making process during selection of miners. Performance of this model is enhanced via an Elephant Herding Optimization (EHO) Model, which assists in incremental improvements in QoS via tuning the Deep Dyna Q Parameter sets. The model was evaluated using simulations of real-world IoT networks and compared with existing QoS-aware security models. The results demonstrate that bioinspired Deep Dyna Q outperforms other models in terms of QoS metrics such as throughput, delay, and packet delivery performance levels. The proposed model has significant implications for the development of efficient and reliable Blockchain-based IoT networks.

Deep, Dyna, Elephant, Herding, Optimization, Blockchain, Security, QoS, Levels