The cuckoo search (CS) technique is applying to discover the optimal route from source to destination. The main objective of this work is to offer suitable solutions for getting better the optimal routing and communicating the data via reliable sensor nodes. This CS optimization method not capable for managing the diversity of the solutions. To solve this issue, we use the CS technique to hybridize it with the hill climbing (HC) technique to minimize the probability of early convergence. This approach introduces a fusion of CS and HC techniques (CSHC) based optimal forwarder selection and detect the Intrusion in wireless sensor network (WSN) . Here, a Bayesian thresholding method is predict the received signal strength and link reliability parameter for identifying intrusion in the network. The hill - climbing technique is able to attain the best solutions in a smaller period than other local search techniques. In CSHC, the optimal forwarderis selection by fitness function. This fitness function is computed based on sensor node lifetime, sensor link reliability, and buffer availability. In this app roach, the experimental results suggest that the CSHC for improving 35% throughput and minimizes the 23.52% packet lossescompared to the baseline approaches.

Clustering <span>is a significant idea for extending the scalability and enhancing the energy in the mobile ad-hoc network (MANET). In addition, the clustering concept is used to diminishes the cost of communication. The re-clustering procedure makes expensive, and frequent re-clustering procedure makes extra routing overhead and extra energy utilization. To solve these issues, received signal strength indication (RSSI) based clustering and aggregating data (RCAD) using Q-learning in MANET is proposed. In this approach, we build the clusters by node RSSI. The fuzzy logic system (FLS) is used to select the cluster head (CH) by the node mobility and node utilization energy. Q-learning-based data-aggregation for improving mobile node routing efficiency in MANET. Here, we can find an optimum next-hop node utilizing their Q-values established on the rewards (RD). Since the RD rule is used to decide the best solution for the Q-learning technique. This RD is computed by present bandwidth (PB), present energy (PE), present packet delivery (PDD), and hop count (HC) parameter for selecting the data aggregator from sender to receiver. The experimental outcomes illustrate that the RCAD approach increases 155 CH round and raises 24% cluster lifetime in the MANET.</span>