Jaya Hunger Game Search Enabled Energy Efficient Optimal Resource Allocation in Internet of Things and Energy Prediction Using Deep Long Short Term Memory

For the past few years, the Internet of Things (IoT) has received considerable attention and developments among researchers to achieve improved resource utilization. For data transmission among heterogeneous components, IoT needs a powerful communication network. So far, some methodologies have been developed for optimal resource allocation, but such techniques are very complex and generally time-consuming models. To counter such issues, this paper proposes a novel hybrid optimization and deep learning framework, Jaya-Hungry Game Search (Jaya-HGS) with Deep LSTM, for energy-efficient resource allocation in IoT networks. Here, the Jaya-HGS is the combination of the Hunger Game Search (HGS) and the Jaya algorithm. The functions considered here for resource allocation are predicted energy, makespan, communication cost, and execution time, in which energy is predicted utilizing Deep Long Short Term Memory (Deep LSTM). In addition, the designed Jaya-HGS has gained maximum energy of 0.121 J, minimum execution time of 1.821 s and maximum throughput of 0.911 Mbps. For 20 rounds, the Jaya-HGS demonstrated improvements of 60.24%, 51.88%, 19.46%, 14.33%, 13.65%, 10.58%, and 5.46% over Transformer-based Deep Reinforcement Learning (TDRL), Opposition-based Learning-Moth-Flame Optimization (OBLMFO), Deep Recurrent Q -learning Networks (DRQNs), Graph Neural Network (GNN), Energy-efficient, Congestion Resource allocation and Routing protocol (ECRR), AI (Artificial Intelligence)-driven Collaborative Dynamic Resource Allocation (ACDRA), and the Deep Learning (DL) routing protocol, respectively, based on energy.